Blog Classic - FindAI Jobs, training and advice

31May

HR Data Scientist at NN Group – The Hague – Haagse Poort – High Rise

As an HR Data Scientist at NN Group, you will be working on the cutting edge of data analytics at the largest insurance company in the Benelux. You will help us answer important HR questions so that our workforce can focus on caring for what matters most to our customers and stakeholders.

What you are going to do

This role will place you at the intersection of HR, data, generative AI and other advanced analytics at a company that is leading in the field of data science. On a typical day, you might be investigating an equal pay analysis method to ensure gender parity. Or analysing employee skills to create a bigger pool to choose from for internal vacancies. Or conducting a cluster analysis to pinpoint frequently asked questions so that we can streamline our HR portal and better serve our employees. Furthermore, you will:

Analyse and solve HR-related questions using innovative data science and analysis techniques to improve the employee experience and data-driven decision-making in HR
Manage the full data science cycle, from identifying an issue and retrieving data from our global HR data platform to presenting the answers in an easily digestible format
Support stakeholders in transforming business challenges into data-driven questions that can be resolved through data analysis
Explain and visualise your findings to stakeholders such as management and HR directors and advise on the next steps
Promote data literacy throughout the HR department by clearly explaining what the data on the dashboard means, especially to colleagues who are not that data-minded

What we offer you

NN invests in an inclusive, inspiring work environment and in skills and competences for the future. We match this with employee benefits that are in line with what is needed today and in the future. This way, we offer our employees the opportunity to get the best out of themselves. We offer you:

Salary between €3566 – € 5095 depending on your knowledge and experience
13th month and holiday allowance are paid with your monthly salary
27 vacation days for a 5-day working week and one Diversity Day
A modern pension administered by BeFrank
Plenty of training and learning opportunities
NS Business Card 2nd class, which gives you unlimited travel, also privately. Do you prefer to travel with your own transport? Then you can declare the kilometres travelled
Allowances for setting up your home office and for internet use

Who you are

You are ready to bring innovative ideas to the table that will help us solve our HR questions, while also being eager to learn and grow. With your innate patience and excellent communication skills, you can easily explain, convince and motivate others to become more data-savvy and base more of their decisions on data. You also have:

A Bachelor degree in data science or mathematics with experience in HR or psychology, or a Bachelor degree in HR or social science with experience in advanced data science
3 year working experience in data analytics
Proficient in python, SQL or a similar open programming language
Preferably experience with large language models or advanced text analytics
Preferably experience with a visualisation tool such as Power BI or Tableau and the Azure data platform

Who you will work with

You will be part of a small, enthusiastic, international team of HR data scientists and HR reporting specialists who work in an informal atmosphere. In this role, you will have a lot of freedom to work independently on your focus area. You will also be required to collaborate with HR colleagues within several areas and with IT and data science departments. Along with 40 other data scientists, you will attend get-togethers and knowledge sharing of the Data Science Guild. You will also enjoy regular meetups with direct colleagues on our office Days.

Any questions?

If you have any questions, you can reach out via phone or WhatsApp to the recruiter via Ties Herpers via

ti**********@nn******.com

.

Source link

31May

DSPy & The Principle Of Assertions | by Cobus Greyling | May, 2024

The principle of Language Model (LM) Assertions is implemented into the DSPy programming framework.

The objective is to make programs more steerable, reliable and accurate in guiding and placing a framework in place for the LLM output.

According to the study, in four different text generation tests LM Assertions not only helped Generative AI Apps follow rules better but also improved task results, meeting rules up to 164% more often and generating up to 37% better responses.

When an assertion constraint fails, the pipeline can back-track and retry the failing module. LM Assertions provide feedback on retry attempts; they inject erring outputs and error messages to the prompt to introspectively self-refine outputs.

There are two types of assertions, hard and soft.

Hard Assertions represent critical conditions that, when violated after a maximum number of retries, cause the LM pipeline to halt, if so defined, signalling a non-negotiable breach of requirements.

On the other hand, suggestions denote desirable but non-essential properties; their violation triggers the self-refinement process, but exceeding a maximum number of retries does not halt the pipeline. Instead, the pipeline continues to execute the next module.

DSPy Assert

The use of dspy.Assert is recommended during the development stage as checkers or scanners to ensure the LM behaves as expected. Hence a very descriptive way of identifying and addressing errors early in the development cycle.

Below is a basic example on how to formulate an Assert and Suggest.

dspy.Assert(your_validation_fn(model_outputs), "your feedback message", target_module="YourDSPyModuleSignature")dspy.Suggest(your_validation_fn(model_outputs), "your feedback message", target_module="YourDSPyModuleSignature")

When the assertion criteria is not met, resulting in a failure, dspy.Assert conducts a sophisticated retry mechanism, allowing the pipeline to adjust. Hence the program or pipeline is not necessarily terminated.

On an Assert failing, the pipeline transitions to a special retry state, allowing it to reattempt a failing LM call while being aware of its previous attempts and the error message raised.

After a maximum number of self-refinement attempts, if the assertion still fails, the pipeline transitions to an error state and raises an AssertionError, terminating the pipeline.

This enables Assert to be much more powerful than conventional assert statements, leveraging the LM to conduct retries and adjustments before concluding that an error is irrecoverable.

DSPy Suggest

dspy.Suggest is best utilised as helpers during the evaluation phase, offering guidance and potential corrections without halting the pipeline.

dspy.Suggest(len(unfaithful_pairs) == 0, f"Make sure your output is based on the following context: '{context}'.", target_module=GenerateCitedParagraph)

In contrast to asserts, suggest statements provide gentler recommendations rather than strict enforcement of conditions.

These suggestions guide the LM pipeline towards desired outcomes in specific domains. If a Suggest condition isn’t met, similar to Assert, the pipeline enters a special retry state, enabling retries of the LM call and self-refinement.

However, if the suggestion consistently fails after multiple attempts at self-refinement, the pipeline logs a warning message called SuggestionError and continues execution.

This flexibility allows the pipeline to adapt its behaviour based on suggestions while remaining resilient to less-than-optimal states or heuristic computational checks.

LM Assertions, a novel programming construct designed to enforce user-specified properties on LM outputs within a pipeline. — Source

Considering the image below, two Suggests are made, as apposed to Asserts. The first suggestions a query length, and the second creating a unique value.

dspy.Suggest(
len(query)     "Query should be short and less than 100 characters",
)dspy.Suggest(
validate_query_distinction_local(prev_queries, query),
"Query should be distinct from: "
+ "; ".join(f"{i+1}) {q}" for i, q in enumerate(prev_queries)),
)

I believe much can be gleaned from this implementation of guardrails…

The guardrails can be described in natural language and the LLM can be leveraged to self-check its responses.
More complicated statements can be created in Python where values are parsed to perform checks.
The flexibility of describing the guardrails lend a high order of flexibility in what can be set for specific implementations.
The division between assertions and suggestions is beneficial, as it allows for a clearer delineation of checks.
Additionally, the ability to define recourse adds another layer of flexibility and control to the process.
The study’s language primarily revolves around constraining the LLM and defining runtime retry semantics.
This approach also serves as an abstraction layer for self-refinement methods into arbitrary steps for pipelines.

⭐️ Follow me on LinkedIn for updates on Large Language Models ⭐️

I’m currently the Chief Evangelist @ Kore AI. I explore & write about all things at the intersection of AI & language; ranging from LLMs, Chatbots, Voicebots, Development Frameworks, Data-Centric latent spaces & more.

Source link

31May

Data Management Associate at JPMorgan Chase & Co. – Bengaluru, Karnataka, India

As a Data Governance Associate for JP Morgan Chase’s Chief Administrative Office (CAO), you will fulfill a central role in the CAO Data and Analytics Office’s book of work, which includes data governance, data risk management and data strategy. The primary objective of the CAO Data Governance team is to provide the resources, tools and support necessary to monitor, manage and prevent ‘Data Risks’ across a global landscape, lead strategic projects that drive business value, and foster a culture of data ownership throughout the organization.

In this role, you will partner with Subject Matter Experts across the Firm to manage and support the overarching Data Use, Data Privacy, Data Protection risk related obligations. Oversee acceptable use of CAO’s data, ensuring adherence to Data Risk and Data ethics guidelines set forth in the firm’s internal policies, global regulatory and legal requirements. You will also support evaluation, design and improvements to control frameworks which mitigates Data related risk, foster a culture of Data Privacy awareness, and escalate issues to appropriate CAO leadership and governance forum as necessary.

Key responsibilities includes:

Support development, maintenance and oversight of complex, cross functional processes for CAO’s Data Privacy and Protection Risk Management Framework, including Records of Processing (Personal Information (PI) Inventory), Data Privacy and Protection Risk Assessments, Privacy Incident Responses, Privacy Notices, Data Subject Requests, and Data Protection
Support the organization in implementing a Privacy by Design approach to embed privacy into our products, services, applications, systems, or processes at the point they are being conceptualized and developed
Facilitate Data Use Case intake, prioritization and assessment processes to ensure handling and/or sharing of CAO’s data internally or externally is in accordance with Firmwide policy, global laws and/or regulatory requirements
Coordinate holistic review of data use cases with key stakeholders, leadership and data use council, when needed, to drive appropriate decisioning
Prioritize Data Use Cases factoring in key deadlines, project go-live dates, and associated levels of risk, ensuring that cases are handled in a timely manner
Participate on Firmwide workgroups supporting data risk management, maintaining adequate knowledge of applicable policy and standards, technology and tooling and supporting adhoc projects, as needed
Partner with leadership in monthly compliance and program metric reviews, supporting corrective actions where needed, and provide commentary for reporting
Support development of job-aids and training materials that enable data owners and other key partners in identifying, managing and mitigating data related risks
Support other areas of data risk management (Quality, Metadata, Privacy, Storage, Retention & Destruction, Classification), as needed

Required qualifications, capabilities and skills:

3 to 5+ years of financial services industry experience in Data Governance, Data Management, Risk Management, Project Management or Business Analysis role in the past
Strong organizational skills; Detail-oriented and ability to multi-task in a fast paced environment with frequently changing priorities and to meet deadlines under pressure
Team player and collaborative mindset; Ability to work independently; High Emotional Intelligence
Strong communication skills (both written and verbal). Ability to communicate business and technical concepts to both expert and novice audiences, and to all levels of the organization
Ability to run meetings across diverse functions, locations and businesses
Intermediate to advanced knowledge of Microsoft office (Excel, PowerPoint, Visio, Word, SharePoint)

Nice To Haves:

Knowledge of Global Data Privacy Laws (e.g. GDPR, CCPA etc.) and requirements with an ability to relate/ translate privacy, compliance and legal requirements into data governance processes, procedures and controls
Foundational knowledge on Information Security, Data Ethics, AI Governance concepts in relation to AI/ML, LLM, Big Data and other emerging areas
Prior experience with Risk & Control Assessment, Audit or Compliance
Familiarity and/or general use of business intelligence tools (QlikSense, QlikView, Tableau, PowerBI)
Experience with industry standard data management tools (Atlan, Informatica, Collibra, OneTrust) and/or reporting tools (Qlikview, Tableau, PowerBI)
Agile/Scrum experience and associated tools (Confluence, Jira, Kanban)
CIPP/CIPM/CISM Certification

JPMorgan Chase & Co., one of the oldest financial institutions, offers innovative financial solutions to millions of consumers, small businesses and many of the world’s most prominent corporate, institutional and government clients under the J.P. Morgan and Chase brands. Our history spans over 200 years and today we are a leader in investment banking, consumer and small business banking, commercial banking, financial transaction processing and asset management.

We recognize that our people are our strength and the diverse talents they bring to our global workforce are directly linked to our success. We are an equal opportunity employer and place a high value on diversity and inclusion at our company. We do not discriminate on the basis of any protected attribute, including race, religion, color, national origin, gender, sexual orientation, gender identity, gender expression, age, marital or veteran status, pregnancy or disability, or any other basis protected under applicable law. We also make reasonable accommodations for applicants’ and employees’ religious practices and beliefs, as well as mental health or physical disability needs. Visit our FAQs for more information about requesting an accommodation.

Source link

30May

Statistical Programmer (Remote) at Mayo Clinic – Rochester, MN, United States

As a Statistical Programmer, the candidate will support Mayo Clinic research by providing high-value analytical collaborations in the form of data manipulations, analysis and reporting. Current Statistical Programmer needs are focused in various clinical departments.

Key roles and responsibilities include the following:

Interact and collaborate within multidisciplinary teams, under the direction of project team leads
Familiarity with programming languages and techniques (e.g., SAS, R)
Contribute to project deliverables by critically reviewing and preparing data for data processing, analysis, and/or summarization
Demonstrate strengths in organization, documentation, logical and systematic thinking, written and oral communication
Perform tasks with strong problem-solving skills, with high attention to detail focusing on delivering high quality results
Manage multiple tasks with concurrent deadlines
Work independently and remotely as well as in a team environment
Seek educational opportunities and share knowledge within teams to enhance professional development

Note: Visa sponsorship is not available for this position. Must be a US citizen, permanent resident, refugee or asylee.

A minimum of a Bachelor’s degree with a major in statistics, biostatistics, bioinformatics, mathematics, computer science, data science, or quantitative degree relevant to the current needs, is required.

A complete application includes a cover letter, resume, and academic transcripts.

Preferred qualifications include the following:

Applied experience and/or coursework in programming (e.g., SAS, R, Python), data capture systems, data management, and statistical analysis
Commitment to customer service
Basic knowledge of human physiology and/or medical terminology
Interest in professional growth and continuing education
GPA of 3.0 or greater

Why Mayo Clinic

Mayo Clinic is top-ranked in more specialties than any other care provider according to U.S. News & World Report. As we work together to put the needs of the patient first, we are also dedicated to our employees, investing in competitive compensation and comprehensive benefit plans – to take care of you and your family, now and in the future. And with continuing education and advancement opportunities at every turn, you can build a long, successful career with Mayo Clinic. You’ll thrive in an environment that supports innovation, is committed to ending racism and supporting diversity, equity and inclusion, and provides the resources you need to succeed.

Source link

30May

ML/AI Engineer / NLP Expert – Custom LLM Development (x/f/m) at HelloBetter – Remote

Join us to create the world’s best AI Therapist that will make conversational mental health care universally accessible, affordable and effective to anyone with internet access.

HelloBetter is a leading company in the emerging field of digital therapeutics (DTx) for mental health conditions. We are also a global pioneer in digital mental health research. Our scientifically validated DTx treatments cover a broad range of mental health disorders, including depression, panic, stress and burnout, insomnia, vaginismus, chronic pain and diabetes, alcohol addiction as well as two prevention programmes.

Founded in 2015 by Psychologists & global pioneers in digital mental health
+€25M total funding
+15 years of research history
€10M in research grants
+100k patients treated
+130 team from +19 countries

For more information, check our HelloBetter Handbook.

We recently entered a new era of technology, marked by the introduction of advanced AI systems using transformer architectures, alongside a rapid growth in computational capacities, potentially having an impact that could exceed that of the internet’s invention three decades ago.

With the growth in the number of patients using our products, HelloBetter has gained invaluable insights into how patients benefit from digital mental health programmes. Now, we deploy AI in order to leverage the insights we gained in +15 years of research, having treated +100k treated patients.

From the beginning we defined our vision as making safe, evidence-based, high-quality digital mental health care accessible to whoever might benefit from it. LLMs as a technology allow us to deliver on this vision by offering AI talk therapy at unprecedented levels of interactivity, personalisation and scale.

Envision an app that provides real talk therapy via chat, voice or video interaction, available 24/7 worldwide, empowering millions of people, especially in low- to middle-income countries with underdeveloped healthcare to improve their mental wellbeing in a self-determined way.

Ultimately, we aim to deliver on the promise that mental health is human right: Do you want to be part of that mission?

YOUR OBJECTIVES

As the leading expert, you and our team of Machine Learning, LLM, and NLP experts, as well as world renowned psychologists and researchers will pioneer the development of our domain-specific Large Language Model (LLM) for mental health care:

Design and architect an AI system with multiple large language models, including a main conversational agent and supplementary models for contextual support.
Steer the fine-tuning of language models on specific healthcare domains to ensure high relevance and accuracy in model responses.
Implement a RAG system to augment the conversational agent with relevant external knowledge and context.
Develop and manage datasets, ensuring data quality and compliance with data privacy regulations.
Establish performance metrics to evaluate model effectiveness and system efficiency.
Collaborate with cross-functional teams, including data scientists, data engineers, software developers, and product managers, to integrate AI models into existing platforms.
Stay updated with the latest advancements in AI, machine learning, and related technologies, applying this knowledge to improve system capabilities.

YOUR PROFILE

Must-Haves:

Bachelor’s, Master’s or PhD in Computer Science, Artificial Intelligence, Machine Learning, or a related field.
2+ years experience building and maintaining NLP/Large Language Models on scale
Strong experience in developing and deploying large language models and NLP applications.
Strong experience in fine-tuning LLMs
Proficiency in machine learning frameworks (e.g. TensorFlow, PyTorch).
Experience with retrieval-augmented techniques and knowledge-intensive NLP tasks.
Excellent programming skills in Python and familiarity with software development best practices.
Demonstrated ability to design and implement complex AI systems.
Excellent communication skills.

Nice-to-Haves:

Experience with cloud services (AWS, Azure) and managing scalable AI solutions.
Knowledge of data privacy laws and ethical AI use especially in the healthcare domain.
Publication record in AI/ML fields or active participation in relevant professional communities.

WHY US?

Meaningfulness

Mental health is a human right: we help thousands of people each month who struggle with depression, stress, insomnia, burnout, and other mental health issues

Research & Evidence

We have a unique product and are at the forefront of research in digital health applications
The effectiveness of our product is continuously evaluated and efficacy studies have been published in international and high-impact journals since 2014
Data is of great importance to us and we are transparent about our strategy, goals and results.

Growth

As pioneers in the development of applications for various mental health conditions, we are at the forefront of innovation
We operate in an extremely exciting and emerging market
Annual training budget of 1,000 euros – we place great emphasis on the personal growth of our employees and actively support their development

Remote-First

Remote-first culture – we hire globally, considering a time window of +/- 4.5 hours CET (This position is part of a research project supported by the state of Brandenburg. Residency around Brandenburg/Berlin is therefore necessary.)
Use of our offices in Berlin and Hamburg if you prefer to work on-site
Relocation option and support

Diversity & Inclusion

Fair and equal treatment are the standards of our Anti-Harassment Policy
Flexible working hours – shape your own day
Company language English, with a strong emphasis on inclusive language
Transparent salary bands
Additional 10 paid leave days for non-birth parents after the birth or adoption of a child

Other Benefits

28 vacation days + additional holidays for public holidays that fall on weekends
Tenure based paid time off – up to three additional days
Permanent employment contract
Attractive VSOP (Virtual Stock Option Plan) for all employees
Tax-deductible pension plan with an above-average employer contribution
Free or subsidized fitness memberships
Regular team events

HelloBetter is an equal opportunity employer and encourages applicants of any national origin, gender, sexual orientation, religious background, gender identity, and people with disabilities to apply.

INTERVIEW PROCESS

Introduction Chat (45 mins)
- Participants: Florian (Talent Partner)
- Objective: Introduce HelloBetter, discuss the mission and team, share your experience, understand the company’s vision, and address any questions about the role and organization.
Technical Interview (60 mins)
- Participants: Vincent (Lead Machine Learning Engineer), Amit (CTO)
- Objective: Assess your technical knowledge and experience relevant to LLMs, NLP, and AI. Dive into your technical experience with AI/ML, understanding of AI technologies, and problem-solving capabilities in an AI context.
Technical Challenge (3 hours)
- Objective: Demonstrate your technical skills and how you would lead development. Solve a given problem, develop a strategy, and present a solution including defining the scope, integrating AI technologies, and addressing challenges.
Technical Challenge Q&A (60 mins)
- Participants: Vincent (Lead Machine Learning Engineer), Amit (CTO)
- Objective: Present your technical challenge solution, receive feedback, discuss your decision-making process, and delve into deeper technical and strategic discussions.
Hiring Manager Interview (60 mins)
- Participants: Amit (CTO and future manager)
- Objective: Discuss your motivation, career aspirations, company culture, and strategy. Understand your strategic vision for the AI Therapist project, align with HelloBetter’s goals, and explore your leadership and team collaboration skills.
Offer Talk
- Participants: Amit (CTO and future manager)
- Objective: Present our offer proposal, discuss your start date, relocation, onboarding, and your first 90 days.

ABOUT US

HelloBetter was founded under the name GET.ON Institut für Online Gesundheitstrainings GmbH by internationally recognized researchers and psychologists in 2015. In more than 30 randomized-controlled studies, HelloBetter has developed and evaluated 10 online programs from eight specific problem areas in cooperation with Leuphana University, Friedrich-Alexander University Erlangen-Nuremberg, the Free University of Amsterdam, Harvard University and other partners. These programs serve both the prevention and treatment of classic mental illnesses such as depression, anxiety or panic disorders, but also cover topics such as vaginismus or chronic pain. This gives HelloBetter the broadest study base of all providers.

In the meantime, six therapy programs from HelloBetter have been approved by the German Federal Institute for Drugs and Medical Devices (BfArM) as digital health applications and are available free of charge on prescription for all adults with health insurance in Germany. The work of the expert team has been published in international journals (including The Journal of the American Medical Association) and recognized with various awards (including, Wilhelm Exner Award in Psychology 2016, EFPA Comenius Award 2017, Digital Health Award 2018, EU Compass Good Practice 2018 in Mental Health).

The company is based in Berlin and Hamburg and employs more than 130 people.

For more information, check our HelloBetter Handbook.

Source link

30May

Comparing LLM Agents to Chains: Differences, Advantages & Disadvantages | by Cobus Greyling | May, 2024

RPA Approach

Prompt chaining can be utilised in Robotic Process Automation (RPA) implementations. In the context of RPA, prompt chaining can involve a series of prompts given to an AI model or a bot to guide it through the steps of a particular task or process automation.

By incorporating prompt chaining into RPA implementations, organisations can enhance the effectiveness, adaptability, and transparency of their automation efforts, ultimately improving efficiency and reducing operational costs.

Human In The Loop

Prompt chaining is ideal for involving humans and by default chains are a dialog turn based conversational UI where the dialog or flow is moved forward based on user input.

There are instances where chains do not depend on user input, and these implementations are normally referred to as prompt pipelines.

Agents can also have a tool for human interaction, the HITL tool are ideal when the agent reaches a point where existing tools do not suffice for the query, and then the Human-In-The-Loop Tool can be used to reach out to a human for input.

Managing Cost

Managing costs is more feasible with a chained approach compared to an agent approach. One method to mitigate cost barriers is by self-hosting the core LLM infrastructure, reducing the significance of the number of requests made to the LLM in terms of cost.

Optimising Latency

Optimising latency through self-hosted local LLMs involve hosting the language model infrastructure locally, which reduces the time it takes for data to travel between the user’s system and the model. This localisation minimises network latency, resulting in faster response times and improved overall performance.

LLM Choose Action Sequence

LLMs can choose action sequences for agents by employing a sequence generation mechanism. This involves the LLM generating a series of actions based on the input provided to it. These actions can be determined through a variety of methods such as reinforcement learning, supervised learning, or rule-based approaches.

Seamless Tool Introduction

With autonomous agents, agent tools can be introduced seamlessly to update and enhance the agent capabilities.

Design Canvas Approach

A prompt chaining design canvas IDE (Integrated Development Environment) would provide a visual interface for creating, editing, and managing prompt chains. Here’s a conceptual outline of what features such an IDE might include: Visual Prompt Editor, Prompt Library, Connection Management, Variable Management, Preview and Testing, etc.

Overall, a prompt chaining design canvas IDE would provide a user-friendly environment for designing, implementing, and managing complex conversational flows using a visual, intuitive interface.

No/Low-Code IDEs

Agents are typically pro-code in their development where chains mostly follows a design canvas approach.

Agents often involve a pro-code development approach, where developers write and customise code to define the behaviour and functionality of the agents. Conversely, chains typically follow a design canvas approach, where users design workflows or sequences of actions visually using a graphical interface or canvas. This visual approach simplifies the creation and modification of processes, making it more accessible to users without extensive coding expertise.

I need to add that there are agent IDEs like FlowiseAI, LangFlow, Stack and others.

⭐️ Follow me on LinkedIn for updates on Large Language Models ⭐️

Source link

30May

Does the EU’s MiFIR Review make single-name credit default swaps transparent enough? – European Law Blog

30 May 2024/
By Randy Priem

Blogpost 29/2024

Regulation 2024/791 (“MiFIR Review”) was published in the Official Journal of the European Union on 8 March 2024. This newly adopted legislation requires single-name credit default swaps (CDSs) to be made subject to transparency rules, only however if they reference global systemically important banks (G-SIBS) or those referencing an index comprised of such banks.

In this blog post, I discuss the suitability of the revised transparency requirements for single-name CDSs of the MiFIR Review. On the one hand, it seems that the new requirements are limited in scope as any referencing entity that is not a G-SIB will not be majorly impacted (see, in more detail, my recent working paper). Indeed, CSDs referencing G-SIBS represent only a small fraction of the market: i.e., 8.36% based on the total notional amount traded and 5.68% based on the number of transactions (source: DTCC). It follows that a substantial percentage of the single-name CDS market will not be captured. On the other hand, this post cautions against creating even more far-reaching transparency requirements than those provided for in the MiFIR Review: more transparency could, in practice, be detrimental for financial markets as it could result in higher trade execution costs and volatility and could even discourage dealers from providing liquidity.

Single-name credit default swaps and why they are opaque.

CDSs are financial derivative contracts between two counterparties to ‘swap’ or transfer the risk of default of a borrowing reference entity (i.e., a corporation, bank, or sovereign entity). The buyer of the CDS – also called the ‘protection buyer’ – needs to make a series of payments to the protection seller until the maturity date of the financial instrument, while the seller of the CDS is contractually bound to pay the buyer a compensation in the event of, for example, a debt default of the reference entity. Single-name CDSs are mostly traded in the over-the-counter derivatives markets, typically on confidential, decentralized systems. A disadvantage, however, of over-the-counter derivative markets is that they are typically opaque, in contrast with, for example, listed financial instruments.

Over-the-counter derivative markets have very limited access to pre-trade information (i.e., information such as the bid-ask quotes and order book information before the buy or sell orders are executed) and post-trade information (i.e. data such as prices, volumes, and the notional amount after the trade took place),

In March 2023, three small-to-mid-size US banks (i.e. Silicon Valley Bank, Silvergate Bank, and Signature Bank) ran into financial difficulties with spillovers to Europe where Credit Suisse needed to be taken over by USB. During this financial turmoil, the CDSs of EU banks rose considerably in terms of price and volume. For Deutsche Bank, there were even more than 270 CDS transactions for a total of US 1.1 billion in the week following UBS’s takeover of Credit Suisse. This represented a more than four-fold increase in trade count and a doubling in notional value compared with average volumes of the first ten weeks of the year. The CDS market is namely illiquid with only a few transactions a day for a particular reference entity, so this increase in trading volumes was exceptional. On 28 March 2023, the press reported that regulators had identified that a single CDS transaction referencing Deutsche Bank’s debt of roughly 5 million EUR conducted on 23 March 2023 could have fuelled the dramatic sell-off of equity on 24 March 2023 causing Deutsche Bank’s share price to drop by more than 14 percent.

One of the conclusions drawn by regulators, such as the European Securities and Markets Authority (ESMA), on the 24 March event was that the single-name CDS market is opaque (i.e., very limited pre-trade and post-trade market information), and consequently, subject to a high degree of uncertainty and speculation as to the actual trading activity and its drivers.

The Depository Trust and Clearing Corporation (DTCC) indeed provides post-trade CDS information, but the level of transparency is not very high, given that only aggregated weekly volumes are provided rather than individual prices. Furthermore, only information for the top active instruments are disclosed rather than for all traded instruments. Regarding pre-trade information, trading is conducted mostly through bilateral communication between dealers, who might directly contact a broker to trade or use a trading platform to enter anonymously non-firm quotes. However, even when screen prices are available, they are only indicative, and most dealers will not stand behind their pre-trade indicated price because the actual price the dealer will transact with is entirely subject to bilateral negotiations conducted over the phone or via some electronic exchange. Dealers are free to change the price until the moment the trade is mutually closed. The end-users are thus dependent on their dealers and sometimes do not even have access to the pre-trade information because they have to rely on third-party vendors and services that aggregate data. End-users do not know before the trade which price offered by dealers is the best one and do not know which other parties are willing to pay or to sell at, nor do they have comparable real-time prices against which to compare the price of their particular trade.

New transparency requirements in the MiFIR Review

On 25 November 2021, the European Commission published a proposal to amend Regulation No 600/2014 on markets in financial instruments (MiFIR) as regards enhancing market data transparency, removing obstacles to the emergence of a consolidated tape, optimizing trading obligations, and prohibiting receiving payments for forwarding client orders. This initiative was one of a series of measures to implement the Capital Markets Union (CMU) in Europe to empower investors – in particular, smaller and retail investors – by enabling them to better access market data and by making EU market infrastructures more robust. To foster a true and efficient single market for trading, the Commission was of the view that the transparency and availability of market data had to be improved.

The proposal implemented the view of ESMA that the transparency regime that was in place earlier was too complicated and not always effective in ensuring transparency for market participants. For single-name CDSs, the large majority of CDSs are indeed traded over the counter where the level of pre-trade transparency is low. This is because pre-trade requirements only apply to market operators and investment firms operating trading venues. Even for CDSs traded on a trading venue, there is a possibility to obtain a waiver as they do not fall under the trading obligation and are considered illiquid financial instruments. Because of their illiquidity, the large majority of listed single-name CDSs can also benefit from post-trade deferrals where information could even be disclosed only after four weeks.

Regulation (EU) 2024/791 (“MiFIR Review”) was finally approved on 28 February 2024 and entered into force on 28 March 2024. Article 8(a) of the MiFIR Review now requires as pre-trade transparency requirement that when applying a central limit order book or a periodic auction trading system, market operators and investment firms operating a multilateral trading facility or organized trading facility have to make public the current bid and offer prices, and the depth of trading interest at those prices for single-name CDSs that reference a G-SIB and that are centrally cleared. A similar requirement is now there for CDSs that reference an index comprising global systemically important banks and that are centrally cleared. Hence, under the new MiFIR Review, CDSs referencing G-SIBS are subject to transparency requirements only when they are centrally cleared. Such CDSs are, however, not subject to any clearing obligation provided for in the European Market Infrastructure Regulation (Regulation No 648/2012 “EMIR”). This means that data on single-name CDSs referencing G-SIBS that are not cleared or CDSs referencing other entities do not need to be made transparent.

Regarding post-trade transparency, Article 10 of the MiFIR Review requires that market operators and investment firms operating a trading venue have to make public the price, volume, and time of the transactions executed in respect of bonds, structured finance products, and emission allowances traded on a trading venue. For the transactions executed in respect of exchange-traded derivatives and the over-the-counter derivatives referred to in the pre-trade transparency requirements (see above), the information has to be made available as close to real-time as technically possible. The EU co-legislators are further of the view that the duration of deferrals has to be determined utilizing regulatory technical standards, based on the size of the transaction and liquidity of the class of derivatives. Article 11 of the MiFIR Review states that the arrangements for deferred publication will have to be organized by five categories of transactions related to a class of exchange-traded derivatives or of over-the-counter derivatives referred to in the pre-trade transparency requirements. ESMA will thus need to determine which classes are considered liquid or illiquid, and above which size of transaction and for which duration it should be possible to defer the publication of details of the transaction.

Besides the pre- and post-trade transparency requirements for market operators and investment firms operating a trading venue, the MiFIR Review also focuses on the design and implementation of a consolidated tape. This consolidated tape is a centralized database meant to provide a comprehensive overview of market data, namely on prices and volumes of securities traded throughout the Union across a multitude of trading venues. According to Article 22a, trade repositories and Approved Publication Arrangements (APAs) will need to provide data to the consolidated tape provider (CTP). The MiFIR Review is then also more specific on the information that has to be made public by an APA concerning over-the-counter derivatives, which will flow into the consolidated tapes. Where Articles 8, 10 and 11 of MiFIR before referred to ‘derivatives traded on a trading venue’, the MiFIR Review no longer uses this wording with respect to derivatives and refers to ‘OTC derivatives as referred to in Article 8a’, being those subject to the pre-trade transparency requirements. This incorporates again those single-name CDSs that reference a G-SIB and that are centrally cleared, or CDSs that reference an index comprising G-SIBs and that are centrally cleared. Similarly as for the pre-trade and post-trade transparency, data on single-name CDSs referencing G-SIBS that are not cleared or CDSs referencing other reference entities do not need to be made transparent.

Do we want even more transparency?

The MiFIR Review’s revised transparency requirements for single-name CDSs are not very far-reaching, given that CDSs referencing to reference entities that are not a G-SIB are not majorly impacted. Given that CSDs referencing G-SIBS represent only a small fraction of the market (see introduction above), a substantial percentage of CDSs is not captured by the MiFIR Review. In addition, single-name CSDs referencing G-SIBS that are not centrally cleared are also not affected. As there is no clearing obligation on CDSs because they are not sufficiently liquid, a large fraction will not be impacted or can continue to benefit from pre-trade transparency waivers or post-trade deferrals. This entails that a large fraction of the entire CDS market will thus not be affected by the MiFIR Review.

Nevertheless, I argue that even more severe transparency requirements than those foreseen by the MiFIR Review might not necessarily be beneficial for financial markets. Too much transparency can be detrimental to financial markets as it might result in higher trade execution costs and volatility and could even discourage dealers from providing liquidity. In a market, in which there are few buyers and sellers ready and willing to trade continuously, asking for more transparency could lead to even less liquidity as the limited number of liquidity providers would be obliged to make their trading strategies available, giving incentives to trade even less. A total lack of transparency might thus be undesirable to avoid market manipulation or from an investor protection point of view, but full transparency on an illiquid CDS market might dissuade traders even more from trading. The EU’s newly adopted MiFIR Review thus seems to strike an appropriate balance between reducing the level of opaqueness while not harming liquidity.

Source link

30May

Applied Scientist at Dialpad – Vancouver, Canada

About Dialpad

Work Beautifully

Dialpad is the leading Ai-Powered Customer Intelligence Platform that is transforming how the world works together. Based on 4 billion minutes of analyzed voice and meetings data and growing, we have designed one, beautiful workspace that seamlessly combines the most advanced Ai Contact Center, Ai Sales, Ai Voice, and Ai Meetings with Ai Messaging. More than 30,000 innovative brands and millions of people use Dialpad to unlock productivity, collaboration, and customer satisfaction with real-time Ai insights. With initial funding and leadership from Google and leading venture capitalists such as ICONIQ and Andreessen Horowitz, Dialpad has over $200M in ARR and is one of the fastest growing Ai companies in the world.

About the team

Our NLP team at Dialpad is a diverse and vibrant group of scientists who come from varied fields such as linguistics, computational linguistics, machine learning, computer science, material science, biology, and quantum physics. We prioritize innovation and creativity in our pursuit to push the boundaries of Natural Language Processing and Artificial Intelligence.

We deliver natural language understanding (NLU) features for transcribed speech and typed inputs. The major focus of the team currently is enhancing DialpadGPT, our in-house LLM specifically designed for the domain of business communication and delivering features that use DialpadGPT. Beyond the technical skills, we’re a team that values collaboration, continuous learning, and the application of diverse perspectives to solve complex problems.

Together, our goal is to revolutionize business communications, making it more efficient, accurate, and accessible through the power of ASR, NLP and machine learning. Using a variety of tools and technologies we’re driven to continuously improve and refine our products, delivering top-tier solutions to our customers.

Your role

As an Applied Scientist at Dialpad, you’ll be an integral part of our NLP team, working on our in-house LLM, DialpadGPT and other LLMs as we extend our NLP functionality to other languages (French, Italian, German, Spanish, and Dutch). You’ll be involved in all stages of model development, from conceptualization to deployment, focusing on enhancing language understanding in business communications. Collaboration will be key as you work alongside our engineering, design, and product teams to build ground-breaking applications.

If you’re passionate about language, AI, and contributing to a team that’s changing the face of business communications, you’ll find yourself right at home with us.

This position reports to the Manager of the NLP team and has the opportunity to be based in our Vancouver Office or work remotely from anywhere in Canada or the US.

What you’ll do

Develop, implement, and refine state-of-the-art Natural Language Processing and Machine Learning algorithms for Dialpad’s products.
Conduct rigorous evaluation and monitoring of model performances and troubleshoot issues with a keen understanding of resultant business impacts.
Collaborate with cross-functional teams, including engineering, product, and design, to effectively deploy and scale models and algorithms in production.
Contribute to the development of DialpadGPT ensuring high accuracy, cost-effectiveness, and optimal performance in serving at scale.
Submit papers to top-tier academic conferences and journals and contribute to the broader scientific community by reviewing submissions.
Capitalize on your interest in languages to extend our coverage to other languages – Spanish, French, Japanese etc. and to optimize models for other varieties of English.

Skills you’ll bring

Master’s or PhD degree in Linguistics, Computational Linguistics, Computer Science, Machine Learning, or related fields.
Demonstrated experience with machine learning, Python, Pytorch and other relevant tools and technologies.
A broad understanding of current LLM model architectures and techniques for tuning and optimizing LLMs.
Knowledge of one or more of the languages we’ll be adding: French, Italian, German, Spanish, or Dutch.
Strong problem solving and analytical abilities, with the capacity to handle complex technical and analytical problems.
Excellent communication and collaboration skills to effectively work in a multi-disciplinary team.
Familiarity with version control tools like Git for collaborative projects.

Dialpad benefits and perks

Benefits, time-off, and wellness

An apple a day keeps the doctor away—and it doesn’t hurt that we offer flexible time off and great options for medical, dental, and vision plans for all employees. Along with that, employees also receive a monthly stipend to help cover your cell phone bill, home internet bill, and we reimburse for gym membership costs, a variety of wellness events, and more!

Professional development

Dialpad offers reimbursement for expenses related to professional development, up to an annual limit per calendar year.

For exceptional talent based in British Columbia, Canada the target base salary range for this position is posted below. Our salary ranges are determined by role, level, and location. The range displayed on each job posting reflects the target range for new hire salaries for the position. Within the range, individual pay is determined by work location and additional factors, including job-related skills, experience, and relevant education or training. Your recruiter can share more about the specific salary range for your preferred location during the hiring process. Please note that the compensation details listed in British Columbia role postings reflect the base salary only, and do not include bonus, equity, or benefits.

British Columbia, Canada Salary Range$148,700—$173,233 CAD

Culture
We’ve been named a Top Workplace seven times, and a big part of this is because of our collaborative culture that elevates our teammates, celebrates wins, and brings together passion and talent.

Compensation and equity
Teamwork makes the dream work, and Dialpad offers competitive salaries in addition to stock options because each and every Dialer participates in our success.

Diversity, Equity, and Inclusion (DEI) at Dialpad

At Dialpad, we are passionate about Doing the Right Thing. This means we are committed to building a values-driven culture that celebrates identity, inclusion and belonging. As a global company, it’s our responsibility to come together to create a culture where all Dialers can Work Beautifully, Delight Our Users, and Innovate Continuously to bring our world-class product to life.

Every Voice Matters at Dialpad. We build community through our Employee Resource Groups, company-wide celebrations, service days, and a robust internal learning & development program focused on the success of our Dialers.

Don’t meet every single requirement? Studies have shown that women and marginalized groups are less likely to apply to jobs unless they meet every single qualification. At Dialpad we are dedicated to building an inclusive and authentic workplace, so if you’re excited about this role but your past experience doesn’t align perfectly with every qualification in the job description, we encourage you to apply anyways. You may be just the right candidate for this or other roles.

Dialpad is an equal-opportunity employer. We are dedicated to creating a community of inclusion and an environment free from discrimination or harassment.

Source link

29May

Principal Research Engineer – Materials at GKN Aerospace – Westlake, TX, US

Fantastic challenges. Amazing opportunities.

GKN Aerospace is reimagining air travel: going further, faster and greener! Fuelled by great people whose expertise and creativity sets the standards in our industry, we’re inspired by the opportunities to innovate and break boundaries. We’re proud to play a part in protecting the world’s democracies. And we’re committed to putting sustainability at the centre of everything we do, opening up and protecting our planet. With over 16,000 employees across 33 manufacturing sites in 12 countries we serve over 90% of the world’s aircraft and engine manufacturers and achieved sales of £3.35 bn.in 2023. There are no limits to where you can take your career.

Job Summary

As a Principal Research Engineer, you will work in cross-discipline teams to develop differentiating additive manufacturing (AM) technology for aerospace structural applications. Your everyday tasks will be to work within GKN Aerospace’s US Global Technology Center team on development of Laser Wire Directed Energy Deposition (DED) AM. You will take on a wide range of tasks focused on AM material characterization, non-destructive inspection methods, process improvement, and qualification of manufacturing processes while working in collaboration with technology engineers with expertise in controls, process development, post processing, and industrialization of aerospace technologies.
You will support development, integration and implement technical solutions and documentation compliant with all relevant customer, regulatory, and airworthiness requirements, within agreed program, schedule, and manufacturing constraints.

Job Responsibilities

Support AM process development by researching and executing experiments to develop and characterize titanium as well as other aerospace metallic materials for laser wire directed energy deposition.
Lead development of non-destructive evaluation methods of additively produced material that meet aerospace requirements for qualification and certification of flight hardware
Research opportunities for novel methods of new in in situ sensors and analysis methods to improve process and resulting materials as well as predictability and reliability of material properties
Engage with Aerospace OEM customers technical experts and airworthiness authorities in additive and metallic structure to develop technical requirements and project plans for application of additively manufacture components into aircraft programs
Support Technology Readiness Assessments towards technology maturation into production
Stay current with related technical activity within GKN’s global technology team (and through related technology organizations) & ensure effective exchange of results and issues
Develop, implement, and check technical solutions and documentation to ensure they are compliant with all relevant customer, regulatory, and airworthiness requirements, within agreed program, schedule, and manufacturing constraints
Identify technical problems and apply/integrate solutions
Support technical and customer reviews, including the creation and preparation of technical data and presentations
Perform all activities with limited supervision and support other engineers with peer-to-peer reviews
Identify, manage and mitigate risks for assigned work

Job Qualifications

Master of Science degree in Engineering or related technical field

7+ years of relevant experience, or substitute for advanced degrees
Must be able to perform work subject to ITAR regulations and/or program requirements.
Experience with CAD and other engineering software
Experience with additively manufactured metallic materials
Experience in Non-Destructive Evaluation methods for metallic materials

Preferred Qualifications

PhD in a relevant field
Aerospace industry experience
Experience with new product development
Experience with aerospace materials including titanium alloys
Knowledge of manufacturing processes data related to traditional and additive manufacturing including machining, heat treatment, and inspection
Proven ability to review technical documentation
Excellent in co-operation, networking, team-working
Experience creating and maintaining databases
Statistical and data analysis capabilities
Able to analyze and interpret customer requirements.
Able to maintain and report on a budget for assigned activities.
Worked in cross-functional teams

We’ll offer you fantastic challenges and amazing opportunities. This is your chance to be part of an organisation that has proven itself to be at the cutting edge of our industry; and is committed to pushing the boundaries even further. And with some of the best training on offer in the industry, who knows how far you can go?

A Great Place to work needs a Great Way of Working

Everyone is welcome to apply to GKN. We believe that we can only achieve our ambitions through a coming together of diverse minds who enjoy collaborating in an inspirational environment. Through our commitment to diversity, inclusion and belonging and by living our five powerful principles we’ve created a culture where everyone feels welcome to contribute. It’s a culture that won us ‘The Best Workplace Culture Award’. By embracing and celebrating what makes us unique we encourage everyone to bring their full self to work.

We’re also committed to providing an accessible recruitment process, so if you require reasonable adjustments at any stage during our recruitment process please get in touch and let us know.

We are the place where human dreams, plus human endeavour, shape the future of aerospace innovation and technology.

Source link

29May

ETL Automation tester – Assistant Manager at State Street – Hyderabad, India

Role Summary & Role Description

As a QA lead, He/she should would be a team player that can perform end to end testing. This candidate should have a strong understanding of the SDLC and defect life cycle. The candidate would be able to perform analysis of requirements, document a test plan, write and execute manual and automated test cases, and analyze test results.

Expected to perform manual and ETL testing in validating data sources, extraction of data, applying transformation logic, and loading data into target tables, working closely with data engineering functions to ensure a sustainable test approach.
5+ years’ experience in ETL, DWH, Functional & Manual testing.
Well versed with all stages of Software Development Life Cycle (SDLC) and Software Test Life Cycle (STLC).
Experience in SQL scripting
Create ETL Test data for all ETL mapping rules to test the functionality of the logic.
Experience in creating Test Readiness Review(TRR) and Requirement Traceability Matrix (RTM) documents
Experience in preparing the Test plan, Detailed Test cases, writing Test Scripts by decomposing Business Requirements and developing Test Scenarios to support quality deliverables.

Responsible for checking the accuracy and completeness of the ETL process. They authenticate the data sources, check the precision of data extraction, apply the transformation logic, and load data in the data warehouse.
Participate in meetings with business on requirements and functional specifications reviews early in the software development life cycle process.
Execute software testing to include functional testing, system testing, regression testing, and, security and performance testing.

Ensure adherence with all State Street SDLC tollgate requirements for SQA ensuring all proper procedures and controls are in place throughout the software testing process
Analyze business requirements, Agile stories and acceptance criteria and technical specifications provided
Define test plans and design/execute functional, integration and regression test cases
Capture and save test backup documentation
Report and track defects/bugs
Work with Agile team members to research and resolve issues/impediments utilizing proven problem-solving skills

Core/Must have skills

Strong experience in SQL\ETL\DB testing.
Understanding of Data ware house, Data Migration, ETL concepts
Knowledge in UNIX.
Understanding of Test case design, Test data preparation, Test execution, Data validation and Evidence capturing.
Understanding of Quality Assurance procedure
Understanding of project end to end process.
Proficiency in Jira and XRAY test management tool is preferable
Experience in Capital markets is preferable
Experience in Financial investment banking services is preferable.
Strong interpersonal communication skills, both written and verbal
Strong ability to review and understand complex concepts at a detailed level
Ability to multi-task and to take the initiative when needed
Ability to work independently as well as within a team setting
Ability to foster and maintain strong relationships with all partners in the Software Development Lifecycle such as business and development.

Running Automation scripts is a plus but not mandatory

Why this role is important to us

Our technology function, Global Technology Services (GTS), is vital to State Street and is the key enabler for our business to deliver data and insights to our clients. We’re driving the company’s digital transformation and expanding business capabilities using industry best practices and advanced technologies such as cloud, artificial intelligence and robotics process automation.

We offer a collaborative environment where technology skills and innovation are valued in a global organization. We’re looking for top technical talent to join our team and deliver creative technology solutions that help us become an end-to-end, next-generation financial services company.

Join us if you want to grow your technical skills, solve real problems and make your mark on our industry.

About State Street

What we do. State Street is one of the largest custodian banks, asset managers and asset intelligence companies in the world. From technology to product innovation, we’re making our mark on the financial services industry. For more than two centuries, we’ve been helping our clients safeguard and steward the investments of millions of people. We provide investment servicing, data & analytics, investment research & trading and investment management to institutional clients.

Work, Live and Grow. We make all efforts to create a great work environment. Our benefits packages are competitive and comprehensive. Details vary by location, but you may expect generous medical care, insurance and savings plans, among other perks. You’ll have access to flexible Work Programs to help you match your needs. And our wealth of development programs and educational support will help you reach your full potential.

Inclusion, Diversity and Social Responsibility. We truly believe our employees’ diverse backgrounds, experiences and perspectives are a powerful contributor to creating an inclusive environment where everyone can thrive and reach their maximum potential while adding value to both our organization and our clients. We warmly welcome candidates of diverse origin, background, ability, age, sexual orientation, gender identity and personality. Another fundamental value at State Street is active engagement with our communities around the world, both as a partner and a leader. You will have tools to help balance your professional and personal life, paid volunteer days, matching gift programs and access to employee networks that help you stay connected to what matters to you.

State Street is an equal opportunity and affirmative action employer.

Discover more at StateStreet.com/careers

Source link

1 … 274 275 276 277 278 … 310