Staff ML Engineer - Post-Training & Model Foundry

San Francisco Tensor Company; San Francisco (CA); On-site; USD 275,000 - 315,000; Full time

Benefits offered by this job

Relocation assistance

Job summary

SF Tensor in San Francisco is building the fastest GPU compiler and enterprise model-foundry to accelerate AI research. This role owns the post-training pipeline end-to-end, from data curation through RL, distillation and deployment, working with customers and domain experts to turn goals into measurable rewards.

The team values hands-on expertise with SFT, RL (DPO, PPO, etc.) and tools like PyTorch or JAX, and offers relocation assistance to bring you on-site in San Francisco with a competitive

Qualifications

• Experience shipping post-trained models to production.
• Hands-on depth in SFT and RL (DPO, GRPO, PPO or similar).
• Ability to evaluate models and distinguish meaningful signals from noise.

Responsibilities

• Own the post-training pipeline end-to-end from data curation to deployment.
• Design and refine reward functions and RL objectives with domain experts.
• Build an eval harness to inform ship/no-ship decisions within tight windows.
• Structure and generate datasets, including synthetic data pipelines.
• Distill expert models into smaller, deployable variants.
• Identify the critical path and accelerate the time-to-model iteratively.

Skills

Post-training models

SFT & RL

Tools

PyTorch

JAX

Job description

SF Tensor in San Francisco is building the fastest GPU compiler and enterprise model-foundry to accelerate AI research. This role owns the post-training pipeline end-to-end, from data curation through RL, distillation and deployment, working with customers and domain experts to turn goals into measurable rewards.

The team values hands-on expertise with SFT, RL (DPO, PPO, etc.) and tools like PyTorch or JAX, and offers relocation assistance to bring you on-site in San Francisco with a competitive