Note: The job is a remote job and is open to candidates in USA. CATCHES builds physics-backed AI for garment simulation and virtual try-on, used by luxury fashion brands. The role involves working directly with image and video models, focusing on quality, consistency, and performance at production scale to deliver outputs for a global audience.
Responsibilities
- Owning quality, consistency, and performance at production scale
- Fine-tuning, optimising inference, and making deliberate choices about model architecture and tooling
- Evaluating and iterating on visual outputs with a rigorous, metrics-driven approach
Skills
- Strong hands-on experience with diffusion-based image and/or video generation models
- Proven ability to optimise VLMs for performance, quality, and consistency in production environments
- Experience fine-tuning or adapting foundation models for specific domains or outputs
- Solid Python skills and familiarity with model serving and inference infrastructure
- Ability to evaluate and iterate on visual outputs with a rigorous, metrics-driven approach
- Hands-on experience with FLUX and/or LTX
- Familiarity with ControlNet, IP-Adapter, or similar conditioning approaches
- Experience with video generation frameworks and temporal consistency techniques
- Knowledge of garment, fashion, or e-commerce visual generation use cases
- Experience with cloud-based GPU inference (GCP, AWS, or equivalent)
Company Overview