About the role
We are looking for a Technical Artist with 2+ years of experience in the generative AI visual space to help evaluate, improve, and scale Decart's real-time video generative models. You'll combine a sharp creative eye with hands-on technical execution - running scripts, curating data, building benchmarks, and providing fine-grained visual feedback to researchers and product teams.
This is a unique hybrid role at a fast-moving AI lab backed by Sequoia, Benchmark, NVIDIA and Radical Ventures, where you'll work directly on cutting-edge models powering products like real-time virtual try-on and live video transformation. If you're passionate about aesthetics, fluent in GenAI tools, and comfortable in a terminal, this role was built for you.
What you will do:
- Lead quality evaluation efforts for generative video models - building and maintaining benchmarks, curating prompt suites, and comparing model outputs to identify improvements, regressions, and failure modes.
- Running scripts, collecting and organizing image/video datasets, and using AI tools (ChatGPT, Gemini, Cursor) to accelerate evaluation and analysis workflows.
- Working cross-functionally between the research team and product team - translating product priorities into evaluation criteria and providing actionable visual feedback to researchers.
- Eyeballing and deeply analyzing fine-grained differences between model outputs, articulating what's better and why to technical stakeholders.
