Scientist II Machine Learning Engineer (LLM & GenAI Specialist)

present of AI; Paris; Hybrid; EUR 90,000 - 130,000; Full time

Job summary

Owkin in Paris is seeking a seasoned ML engineer to turn research code into reliable production systems for large language models and generative AI. You should have about five years of Python experience, with real production DL work and comfort with PyTorch or TensorFlow, Hugging Face, and cloud tools like SageMaker.

You will help scientists write cleaner, testable code and set up tracking to ensure experiments are repeatable, while coaching researchers on technical tradeoffs and safety in

Job description

This role is about turning research code into reliable machine learning systems, with a strong focus on large language models and generative AI. It asks for roughly five years of Python work, real experience putting deep learning models into production, and comfort with PyTorch or TensorFlow, Hugging Face, and cloud tools like SageMaker. A strong candidate also helps scientists write cleaner, testable code and sets up tracking so experiments can be repeated.

How it fits the AI era

The work sits right inside the AI shift, since it covers building agent style workflows, retrieval systems, and faster model serving and training at scale. AI tools can speed up prototyping and code writing, but the judgment about what is reliable, reproducible, and safe for biomedical use stays with people. Coaching researchers and explaining technical tradeoffs to different audiences also stays human.

How it builds AI experience

Working on a product where an AI scientist supports pharmaceutical research gives hands‑on exposure to agents, language models, and the infrastructure behind them, applied to real biology questions. Anyone wanting to build toward this can practice with open models on Hugging Face, build a small retrieval or agent project on public biomedical data, and try distributed training or fast inference tools. Experience with clinical or omics data would add to the fit.

Worth showing

• Deploying deep learning to production
• LLM, RAG, and agent projects
• PyTorch and Hugging Face depth
• Cloud MLOps with SageMaker
• Testing, CI/CD, and Docker habits

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

About Owkin

Per the company, Owkin builds K Pro, an AI scientist that works alongside pharmaceutical research teams to speed discovery and strategic decisions.