Senior Applied Scientist

What you'll do

• Build reproducible training, evaluation, and deployment loops, including evaluation harnesses and metrics.
• Design reward and feedback signals; investigate reward hacking, specification gaming, distribution drift, and training instability.
• Own data pipelines and automated data flywheels, and collaborate with research and product to ship robust ML systems.

What they ask for

• Excellent Python and ability to write clean, well-tested ML training code.
• Solid grasp of data pipelines, distributed or large-scale training, and experiment tracking.
• Hands-on experience with at least one feedback or learning loop, such as a reward model, evaluation harness, RLHF/RLAIF or active-learning pipeline, retraining pipeline, or fine-tuning with feedback.
• Working knowledge of reward modeling, on-policy versus off-policy tradeoffs, credit assignment, feedback-system failure modes, and evaluation of non-stationary systems; experience shipping and maintaining an ML system in production.

Nice to have

• PhD or MS in RL/ML paired with real production experience.
• Experience at an RLHF, agents, or large-scale ML infrastructure lab or company; familiarity with LLM fine-tuning, evaluation frameworks, or agent orchestration.

Tools

Python, RLHF, RLAIF

AI focus

Build engineering systems for recursive self-improvement, including reinforcement-learning and feedback loops, reward design, evaluation, and safeguards against degradation or objective gaming.