ML Researcher - Posttraining
Posted Updated
Description
You will fine-tune diffusion models at scale and implement post-training techniques including supervised fine-tuning, preference optimization, reinforcement learning, and distillation. You will develop evaluation and reward systems, train VLM and LLM models, coordinate preference-data collection, and work on safety alignment.
Responsibilities
- Fine-tune diffusion models at scale to improve image aesthetics and quality
- Implement supervised fine-tuning, preference optimization, reinforcement learning, distillation, and acceleration techniques
- Design evaluation suites and reward systems for reinforcement learning
- Train VLM reward models
- Train LLMs for prompt expansion
- Coordinate preference-data and model-evaluation collection
- Work on safety alignment for open-source release
- Integrate research advancements into products
Requirements
- Experience post-training diffusion models for image or video generation
- Experience with large-scale model training, inference, and optimization
- Knowledge of LLM and diffusion post-training pipelines
- PPO, GRPO, DPO, OPD, and MOPD
- PyTorch
- Distributed training including FSDP, CP, SP, USP, TP, and EP
- Low-precision training and inference in FP8, NVFP4, and MXFP8
- vLLM and sglang
- Reinforcement learning infrastructure and optimization
- Experience training VLM models
- Knowledge of LLM, VLM, representation learning, and robotics research
Benefits
- Equity packages
- Health insurance
- Dental and vision insurance
- Health FSA accounts
- Long-term disability coverage
- Flexible PTO
- 401k with a 4% company-sponsored match
- Office meals
- Uber transit to and from the office
- Visa sponsorship
Skills
Sign in to read and join the discussion.