10 часов назад
AI Research Engineer (Diffusion Models)
Вакансия из Hirify Global, списка международных tech-компанийДля мэтча и отклика нужен Plus
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
Текст:
TL;DR
AI Research Engineer (Diffusion Models): Finetuning diffusion models at scale and developing posttraining systems for image and video generation with an accent on aesthetics, preference optimization, reinforcement learning, and reward modeling. Focus on designing image-space evaluations, training VLM and LLM reward models, optimizing distributed training and inference, and integrating research advances into AI creative products.
Location: On-site, full-time in North Beach, San Francisco
Company
Krea AI builds next-generation AI creative tools for text, images, video, sound, and 3D, including its Krea 2 foundation model.
What you will do
- Finetune diffusion models at scale to improve image aesthetics and quality.
- Implement supervised finetuning, preference optimization, reinforcement learning, on-policy distillation, and acceleration techniques.
- Design evaluation suites and reward systems for image-space reinforcement learning.
- Train custom VLM reward models and LLMs for prompt expansion.
- Coordinate preference-data collection and model evaluation with data teams and partners.
- Develop safety alignment for open-source model releases and integrate research advances into products.
Requirements
- Proven experience posttraining diffusion models for image or video generation.
- Experience with large-scale model training, inference, and optimization.
- Strong understanding of LLM and diffusion posttraining algorithms, including PPO, GRPO, DPO, OPD, and MOPD.
- Strong proficiency in PyTorch and distributed training paradigms including FSDP, CP, SP, USP, TP, and EP.
- Knowledge of low-precision training and inference with FP8, NVFP4, and MXFP8, plus fast inference engines and RL infrastructure.
- Experience training VLMs, designing concrete evaluations and rewards, and working independently on ambiguous research goals.
Nice to have
- Demonstrated interest in creative fields such as music, design, or visual arts.
- Knowledge of developments in LLMs, VLMs, representation learning, or robotics research.
Culture & Benefits
- Fast iteration, independent execution, and a goal-oriented research environment.
- Competitive salary and equity packages.
- Employees receive covered health, dental, and vision premiums, health FSA accounts, and long-term disability coverage.
- Flexible PTO and a 401(k) with a 4% company-sponsored match.
- Office meals and covered Uber transportation to and from the office.
Hiring process
- No hiring process details are provided.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →