ML Researcher - Posttraining

The role

The ML Researcher will finetune diffusion models at scale to improve image aesthetics and quality. They will implement posttraining techniques and design evaluation suites and reward designs for reinforcement learning.

What you will do

• Finetune diffusion models at scale to improve image aesthetics and quality.
• Implement posttraining techniques including supervised finetuning, preference optimization, reinforcement learning, on-policy distillation, and distillation/acceleration techniques.
• Design comprehensive eval suites and reward designs for the reinforcement learning stage focused on image space.
• Train custom VLM as reward models as part of reward design.
• Train custom LLMs for prompt expansion through finetuning and reinforcement learning.
• Coordinate with data teams and partners to manage collection of preference data and model evaluation results.
• Work on safety alignment of models for open source release.
• Collaborate with AI research and engineering teams to integrate advancements into products.

Required qualifications

• Proven work of posttraining diffusion models for image or video generation.
• Experience with large-scale model training, inference, and optimization.
• Strong understanding of both LLM and diffusion post training pipelines and algorithms such as PPO, GRPO, DPO, OPD, and MOPD.
• Strong proficiency in PyTorch and understanding of its inner workings.
• Strong background in distributed training paradigms such as FSDP, CP, SP, USP, TP, and EP.
• Good knowledge of low precision training / inference in FP8, NVFP4, and MXFP8.
• Good understanding of algorithms and techniques used in fast inference engines such as vLLM and sglang as well as existing RL frameworks in LLM space such as slime, miles, tinker, and verl.
• Understanding of various RL infrastructure and optimization techniques such as async RL, fast weight transfer, pipelining rollouts, managing off policy data.
• Ability to monitor model regression and identify weak areas and turn them into concrete evals and reward design.
• Experience training VLM models.
• Keeping up with the developments in related fields such as LLM, VLM, representation learning, and robotics research.
• Being comfortable working in a goal-oriented research environment.

Benefits

• Generous salary & equity packages
• 100% health insurance premiums covered
• 99% dental/vision insurance premiums covered
• Health FSA accounts
• Long-term disability coverage
• Flexible PTO policy
• 401k with a 4% company-sponsored match
• Breakfast, lunch, dinner covered in the office
• Ubers covered to & from the office

Before you apply

• We're open to sponsoring international visas where we can (e.g., STEM OPT, OPT, H-1B, O-1, E-3).

About the company

Krea is building next-generation AI creative tools to make AI intuitive and controllable for creatives. They have raised over $83M and are backed by investors such as a16z, Bain Capital, and Abstract.