Remote job
Member of Engineering (Post-training)
Job details
About this role
Role overview Join an Applied Research team focused on post-training large language models so they become well-aligned, highly capable systems for coding and software development. The work sits at the intersection of research and engineering, iterating quickly across synthetic data, fine-tuning, and reinforcement learning techniques. It is a full-time, remote role open to candidates located in EMEA or US East Coast time zones, with periodic in-person team gatherings in Paris.
Responsibilities
- Develop and refine post-training methods that turn pre-trained LLMs into capable, aligned coding assistants. - Design, run, and analyze experiments across fine-tuning, synthetic data generation, continual learning, and reinforcement learning techniques. - Build evaluation pipelines that measure coding capability, alignment quality, and real-world usefulness. - Collaborate across research, engineering, infrastructure, and product teams to move ideas from prototype to deployed model improvements. - Stay close to the state of the art in code generation and alignment, applying and contributing new approaches. - Iterate rapidly, questioning existing methods and tooling to improve code quality and model behavior.
Requirements
- Experience working with, or contributing to the state of the art in, topics such as fine-tuning and alignment of LLMs, synthetic data generation, continual learning, RLVR, or code generation. - Comfort operating in a rapidly iterating research and engineering environment. - Strong Linux skills and solid algorithmic foundations. - Programming experience in Python with PyTorch or JAX. - Strong critical thinking skills and willingness to question established code quality and process policies.
Nice to have
- Prior programming experience outside of Python and outside of machine learning contexts.
Benefits and work setup
Fully remote with flexible hours, plus monthly three-day team gatherings in Paris (Monday through Wednesday) and longer annual off-sites. 37 days per year of vacation and holidays, a health insurance allowance covering dependents, 16 weeks of flexible full-pay parental leave, well-being and learning allowances, a home office allowance, company-provided equipment, and a people-first, inclusive culture.