ML Researcher: Diffusion & RL for Creative AI

Krea; San Francisco, Northern (CA, KY); Hybrid; USD 180,000 - 260,000; Full time

Benefits offered by this job

Health & dental insurance

Flexible PTO

401k with company match

Meals in the office

Transit coverage

Visa sponsorship available

Job summary

Krea, a San Francisco-based AI creator tools company, seeks a senior R&D engineer to advance diffusion model posttraining at scale and build reward models. You will collaborate with data teams, researchers, and engineers to improve image aesthetics, align safety, and push capabilities across LLMs and VLMs.

The role requires deep experience in diffusion posttraining, large-scale training, and RL techniques, with a strong PyTorch background and track record of delivering scalable AI systems.

Qualifications

• Proven work of posttraining diffusion models for image or video generation.
• Experience with large-scale model training, inference, and optimization.
• Strong understanding of LLM and diffusion post training pipelines and algorithms such as PPO, GRPO, DPO, OPD, and MOPD.

Responsibilities

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

Skills

Posttraining diffusion

Distributed training

Reinforcement learning (RL)

PyTorch

Model evaluation

Inference optimization

Tools

vLLM

sglang

Job description

Krea, a San Francisco-based AI creator tools company, seeks a senior R&D engineer to advance diffusion model posttraining at scale and build reward models. You will collaborate with data teams, researchers, and engineers to improve image aesthetics, align safety, and push capabilities across LLMs and VLMs.

The role requires deep experience in diffusion posttraining, large-scale training, and RL techniques, with a strong PyTorch background and track record of delivering scalable AI systems.