Research Engineer / Scientist, Post-training & Reinforcement Learning - London
H Company · Hybrid London
CHECKED 8H AGO
Frontier labremote
At a glance
Highlights
- Build foundation LLMs and VLMs that power cutting-edge agentic technology.
- Advance long-horizon instruction following for GUI and computer-use agents.
- Ship model releases that set new SOTA on public leaderboards.
What they offer
- Early-stage opportunity shaping the future of AI at a leading AI startup
- Collaborative, dynamic and multicultural team with world-class AI talent
- Highly competitive salary
- Opportunities for professional growth, continuous learning and career development
Ideal candidate
You are a collaborative, low-ego researcher-engineer energized by hard problems and fast iteration. You care deeply about advancing safe, capable agentic AI and turning research into real impact.
Requirements
- Experience training and improving large language models at different scales in distributed settings
- Strong programming in Python, Rust or similar with solid software engineering fundamentals
- Proficient in deep learning frameworks like PyTorch, JAX, TensorFlow
- Experience with post-training techniques such as SFT, DPO, RLHF/RLVR, reward modelling and distillation
- Experience with offline and online reinforcement learning
- Experience with large-scale distributed training and inference
- remote policy
- Hybrid, 3 days per week in office with travel between offices every 4-6 weeks
Lab profile
About H Company
Pushing superintelligence boundaries with autonomous agentic AI that automates complex multi-step tasks.
What they build
- Autonomous agentic AI systems
- Foundational infrastructure for autonomous AI
Known for
- Post-training
- Model inference and serving
- AI infrastructure
- Safe and responsible AI development
Key facts
- Teams in Paris, London, and the US