Responsibilities
- Collaborate closely with AI engineers, product managers, and content creators to define and execute on visual AI features
- Fine-tune diffusion-based models via full fine-tuning
- Improve and extend our existing image and video generation pipelines by using state-of-the-art technologies and advancements
- Contribute to shared tools and internal libraries that support dataset preparation, experimentation, and quality control
Requirements
- Master’s degree or higher in Computer Vision, Computer Science, Applied Mathematics, or a related technical field
- 2+ years of experience performing full fine-tuning of diffusion-based models for image generation
- Proven track of trained models being deployed in traffic-intensive production environments
- Strong software engineering skills: writing clean, modular, maintainable code; familiarity with version control, testing, and team workflows
- Proficiency in Python and experience with tools and libraries such as PyTorch, Hugging Face Diffusers, Pillow, OpenCV, ComfyUI, or Automatic1111
- Experience working with cloud-based environments (e.g. GCP, AWS) for training, experimentation, and data processing
Nice to Have
- Background in designing intuitive tooling for content creators or technical teams
- Portfolio of visual AI projects or contributions to open-source models or tools
Benefits
- We prefer a B2B contract but we can be flexible, as long as you’re in it for the long haul
- Full-remote (you work from the place that suits you best)
- 4 weeks PTO
- Annual gathering to get to know each other better
- Wellbeing budget up to 200$
- Learning budget
- Company laptop
- GPT-4, Mistral and Hugging Face Pro plan
Work Arrangement
Remote (Worldwide)
Team
Team size: 55. Structure: 55+ people
Additional Information
- Strong communication & collaborative skills (perfectly fluent in English)
- Goal-oriented, ownership and commitment
- Doer mindset - we are moving fast and we need people who can find the right balance between executing, planning and strategy
- Humble - willing to learn, open to feedback
- You are comfortable building products that are based on uncensored models and content
