Machine Learning Engineer — Self-Improving RL Pipelines

Hippocratic AI Inc.; Menlo Park (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Hippocratic AI in Menlo Park, CA is seeking an engineering-focused ML role to build and maintain end-to-end self-improvement loops with strong emphasis on reproducibility and safety. You will design reward signals, evaluation harnesses, and data pipelines for robust, production-ready systems.

Ideal candidates have hands-on RLHF/RLAIF experience, expertise in large-scale training, and a track record of shipping ML systems that stay healthy over time.

Qualifications

• Excellent Python and clean, well-tested ML training code.
• Solid grasp of data pipelines, distributed / large-scale training, and experiment tracking.
• The instinct and skill to debug why a model silently got worse — not just why it crashed.
• Hands-on experience with a feedback or learning loop (RLHF/RLAIF, active‑learning, or data flywheels).
• Experience retraining or continual-learning pipelines using production data or model outputs.

Responsibilities

• Build and maintain the training, evaluation, and deployment loops at the core of the self‑improvement system, with emphasis on reproducibility and reliability.
• Design and implement reward and feedback signals; mitigate reward hacking, specification gaming, and distribution drift.
• Build evaluation harnesses and metrics before models to measure improvement.
• Own data pipelines and automated data flywheels feeding the learning loop.
• Debug subtle model-quality regressions and stabilize non‑stationary training and feedback loops.
• Collaborate with research and product to turn methods into robust, shippable systems.

Skills

Python

ML training code

Data pipelines

Distributed training

Experiment tracking

Debugging models

RLHF / RLAIF

Education

PhD in RL/ML

MS in RL/ML

Tools

LLM fine-tuning tooling

Agent orchestration tools

Job description

Hippocratic AI in Menlo Park, CA is seeking an engineering-focused ML role to build and maintain end-to-end self-improvement loops with strong emphasis on reproducibility and safety. You will design reward signals, evaluation harnesses, and data pipelines for robust, production-ready systems.

Ideal candidates have hands-on RLHF/RLAIF experience, expertise in large-scale training, and a track record of shipping ML systems that stay healthy over time.