- Workplace
- Remote, Hybrid
- Type
- Full-time
- Department
- Education
- Seniority
- Senior
- Closing date
- Today
- Source
- Vincere
Description
Want to work on what comes after today’s LLMs?
Here you’ll join a frontier AI team focused on building towards superintelligence, not just scaling existing models, but improving how they reason, plan, and act over long horizons.
The work centres on training systems that go beyond next-token prediction. Models that can explore, make decisions, use tools, and improve through interaction with complex environments.
A big part of this is post-training. You’ll be working across SFT, reinforcement learning, and agent-based training loops, shaping how models behave rather than just how they score on benchmarks.
There’s also a strong focus on architecture and scale. The team are interested in those who work with mixture-of-experts systems and large models in the 30B+ to 100B+ range, with access to serious compute. This is the kind of environment where you will optimise, experiment, and iterate quickly.
You’ll be joining a small, research-heavy group with real autonomy, backed by a larger organisation with the resources to support ambitious work. Your impact on the development of core foundational work will be meaningful.
High ownership, fast pace, and a clear focus on pushing capability forward.
What you’ll bring:
- Experience working on large models (ideally 30B+ parameters)
- Exposure to MoE or large-scale distributed training
- Background in LLM post-training, RL, or agent systems
- Experience in a frontier lab or similarly ambitious environment
This is the type of environment many people at top labs leave to build themselves. The difference here is you get the backing, compute, and team already in place.
Salary: $350k - $450k Base DOE + Sizeable Stock Options and Bonuses
Location: SF hybrid or US Remote or London
If you’re thinking seriously about where frontier model development is heading next, this is worth a conversation.
All applicants will receive a response.
