Machine Learning Researcher / Engineer (Foundational Models)

Pathway; Paris; À distance; EUR 90 000 - 150 000; Plein temps

Résumé du poste

Join a pioneering AI startup that is redefining the landscape of Live AI. As an R&D Engineer, you will engage in groundbreaking research, working on attention-based models and distributed training. This role offers a unique opportunity to contribute to impactful projects with a flexible GPU budget in a collaborative environment. Ideal candidates will have a strong background in machine learning and deep learning, with a passion for innovation and experimentation. If you're ready to tackle complex challenges and shape the future of AI, this position is for you.

Prestations

Employee Stock Option Plan

Remote work options

Intellectually stimulating environment

Competitive salary

Qualifications

• Experience in machine learning research with a focus on deep learning.
• Hands-on experience with model training and experimentation.

Responsabilités

• Perform distributed model training and improve model architectures.
• Design new tasks and oversee data preparation activities.

Connaissances

Machine Learning Research

Deep Learning

Language Models

Reinforcement Learning

GPU Architecture

Graph Algorithms

Experimentation

Fluent in English

Formation

PhD in Computer Science or related field

Publications in NeurIPS, ICLR, or ICML

Outils

PyTorch

Jax

TensorFlow

Git

CI/CD

Description du poste

About Pathway

Pathway is an enabler for Live AI, allowing organizations to run contextualized ML models connected to ever-changing enterprise data. We are an infrastructure provider delivering an AI framework and working to advance the state-of-the-art.

This is an R&D position in attention-based models.

Pathway is VC-funded, with notable advisors such as Lukasz Kaiser, co-inventor of Transformers. Our CTO has co-authored papers with Geoffrey Hinton and Yoshua Bengio. We recently raised over $10M in seed funding, with exciting developments ahead. Our management team includes growth leaders experienced in scaling companies and building online communities, reaching millions of users.

Our client portfolio includes mobility, IoT data, logs, transactions, NATO, and national postal services. We have a vibrant community around our developer frameworks, with nearly 10,000 stars on GitHub.

Our offices are located in Menlo Park, CA, Paris, France, and Wroclaw, Poland.

The Opportunity

We are seeking 1 or 2 R&D Engineers with a strong record in machine learning research.

This is a highly ambitious foundational project with a flexible GPU budget in the seven-figure range.

You Will

1. Perform distributed model training
2. Improve and adapt model architectures based on experimental results
3. Design new tasks and experiments
4. Optionally oversee data preparation activities of team members

Your work will be crucial to the project's success.

Requirements

Cover letter: We appreciate a brief 2-3 line introduction.

You should meet at least one of the following criteria:

• Published at least one paper at NeurIPS, ICLR, or ICML as lead author or with significant contributions
• Contributed to a newsworthy LLM training effort, such as outperforming benchmarks or creating a best-in-class model using multiple GPUs
• Spent at least 6 months at a leading ML research center (e.g., Google Brain, DeepMind, Apple, Meta, Anthropic, Nvidia, MILA)
• Been an ICPC World Finalist, or medalist at IOI, IMO, or IPhO in high school

You Are

• A deep learning researcher with experience in Language Models and/or Reinforcement Learning (Vision or Robotics ML backgrounds are also welcome)
• Interested in improving foundational architectures and creating benchmarks
• Hands-on experience with experiments and model training (PyTorch, Jax, or TensorFlow)
• Understanding of GPU architecture, memory design, and communication
• Knowledge of graph algorithms
• Familiarity with model monitoring, git, build systems, and CI/CD
• Respectful of others
• Fluent in English

Bonus Points

• Knowledge of distributed training approaches
• Familiarity with Triton
• Successful record in algorithms and data science contests
• A portfolio of code contributions

Why You Should Apply

• Join an intellectually stimulating environment
• Pioneer new challenges in "Live AI" with long sequences and dynamic data
• Be part of an early-stage AI startup committed to impactful research and foundational change

Benefits

• Full-time, permanent contract
• Preferably start in January 2025; positions open until filled
• Competitive six-figure annual salary plus Employee Stock Option Plan
• Remote work with options to meet in offices in Menlo Park, CA, Paris, France, or Wroclaw, Poland; residence in EU, UK, US, or Canada required

If you meet the broad requirements but lack some experience, feel free to reach out to us.