Post-Training ML Research Engineer - Design Judgment Systems

Vizcom; San Francisco (CA); On-site; USD 220,000 - 300,000; Full time

Benefits offered by this job

Employer-sponsored medical coverage

Dental/vision coverage

Equity

Flexible PTO

401(k) with company match

Learning & Development allowance

Parental leave

Weekly tech lunch in SF

Gym stipend

Job summary

Vizcom in San Francisco, CA is seeking a Research Engineer, Post-Training to advance reward and preference models and connect research insights to product features. The role sits between research and production, requiring strong ML engineering and evaluation discipline.

You will collaborate with the product team and engineers to scale models used by designers. The ideal candidate has experience with post-training methods, robust evaluation pipelines, and a focus on real-world design decisions,

Qualifications

• Strong programming and software engineering skills, especially for machine learning systems.
• Experience training or post-training generative models (diffusion/flow).
• Experience designing evaluations and experimentation pipelines you can trust.
• Interest in product-coupled research where insights reach production.

Responsibilities

• Define and execute the post-training roadmap for models from research to production.
• Build reward and preference models from professional design decisions.
• Explore methods like supervised fine-tuning, distillation, preference optimization, and RL.
• Build rigorous evaluations and practices to determine when results are real and scalable.
• Partner with Product to align signals with product needs.

Skills

Programming

Software engineering

ML systems

Post-training models

Evaluation pipelines

Product-focused research

Ambiguous problem solving

Job description

Vizcom in San Francisco, CA is seeking a Research Engineer, Post-Training to advance reward and preference models and connect research insights to product features. The role sits between research and production, requiring strong ML engineering and evaluation discipline.

You will collaborate with the product team and engineers to scale models used by designers. The ideal candidate has experience with post-training methods, robust evaluation pipelines, and a focus on real-world design decisions,