Senior Model Engineer Post-Training and Reinforcement Learning
Posted Updated
Description
You will design rigorous experiments, improve model recipes through data-quality and mixture experiments, define effective reinforcement-learning data and rewards, advance training science, build targeted evaluations, and help scale successful experiments into frontier training runs.
Responsibilities
- Design and run rigorous experiments that produce measurable gains in model capability
- Improve model recipes through data quality, deduplication, and mixture experiments
- Define effective reinforcement learning data and rewards, including grader design and reward-hacking prevention
- Advance training science through architecture, scaling, and post-training experiments
- Build evaluations and targeted tasks that expose model weaknesses
- Partner with engineers to scale promising results from small experiments to frontier training runs
Requirements
- Production-quality programming in Python, PyTorch, JAX, Rust, or C++
- Experience training large models, including pre-training, post-training, fine-tuning, or reinforcement learning training of neural networks with more than 1 billion parameters
- Post-training
- Reinforcement learning
- Reward modeling
- Preference optimization
- RLHF
- DPO
- Model evaluation
- Alignment
Benefits
- Equity
- Comprehensive medical coverage
- Vision coverage
- Dental coverage
- 401(k) retirement plan
- Short-term disability insurance
- Long-term disability insurance
- Life insurance
- Discounts and perks
Skills
Sign in to read and join the discussion.