Research Engineer, Post-Training
VizcomGenerative AI company
San Francisco, United States$220,000–$300,000 USDMid
Radical Ventures
Index Ventures
Unusual Ventures
Basis Set Ventures
Sound Ventures
Nat Friedman
Data & AI
About the role
TL;DR
Build models to learn from professional design decisions and ship them to users.
- •Vizcom is seeking a Research Engineer, Post-Training to build models that understand and learn from professional design decisions.
- •This role bridges research and product, with models shipping to working designers.
- •Key Responsibilities Define and execute the post-training roadmap for models.
- •Build reward and preference models using professional design decisions.
- •Explore and apply methods like supervised fine-tuning, distillation, preference optimization, and reinforcement learning.
- •Build rigorous evaluations and research practices.
- •Partner with Product to shape future data capture.
- •Requirements Strong programming and software engineering skills for ML systems.
- •Hands-on experience training or post-training generative models (diffusion or flow models).
- •Experience with post-training methods (SFT, preference optimization, reward modeling, distillation, RL).
- •Experience designing trustworthy evaluations and experimentation pipelines.
- •Interest in product-coupled research.
Required skills
PythonTensorFlowPyTorchscikit-learnGitLinux
Nice-to-have skills
MLflowKubeflowSageMakerVertex AIONNXJAX
Domain expertise
aiproductivity
Benefits & perks
100% employer-sponsored medical coverage for employees, plus 25% coverage toward dependents, Dental and vision coverage, plus mental health benefits, Meaningful equity ownership, Flexible PTO, 401(k) with employer match, Generous annual Learning & Development allowance, Paid parental leave, Weekly catered lunch at our San Francisco headquarters, Monthly gym membership stipend
Tech stack
PythonTensorFlowPyTorchscikit-learnGitLinux