Research Engineer (Experimental ML Systems)

TL;DR

Research Engineer (Experimental ML Systems): Building experimental systems that study how models generalize, fail, and become misaligned with an accent on designing synthetic RL environments, building research tooling, and running experiments on model behavior. Focus on studying reward hacking, deceptive behavior, goal misgeneralization, and prototyping novel training setups and evaluation harnesses.

Onsite in San Francisco, CA (or other US locations: North Reading MA, Santa Clara CA, Boston MA, United States)

Salary: $150,000-$350,000 (various ranges)

Company

Early-stage AI research lab focused on alignment, interpretability, and reinforcement learning. Revenue-generating.

What you will do

• Design synthetic RL environments for studying model behavior under distribution shift
• Build experimental platforms for generalization, robustness, and alignment research
• Study reward hacking, deceptive behavior, and goal misgeneralization
• Prototype novel training setups and evaluation harnesses
• Develop benchmarks measuring internal consistency, not just outputs
• Rapidly test research hypotheses through code and experiments

Requirements

• Onsite in US locations (San Francisco CA preferred)
• Experimental ML or RL research experience
• Security research, adversarial thinking, or red-teaming
• Building small research prototypes (not production systems)
• Strong engineering ability with curiosity and creativity
• PhD preferred; ability to build novel research systems from scratch

Culture & Benefits

• Highly exploratory work in AI research lab
• Focus on inventing new experimental systems