Research Engineer

Research Engineer, Interpretability Systems - San Francisco, CA (Onsite)

I'm currently working with an early-stage AI research lab founded by former OpenAI and Google researchers focused on one of the most important problems in AI: understanding how models actually work internally.

Their team is building new techniques for interpretability, alignment, reinforcement learning, and representation engineering, developing tools that allow researchers to inspect, measure, and steer model behavior at the activation level. Rather than building applications on top of AI, they're working directly on the underlying mechanisms of intelligence itself.

They're looking for highly technical engineers who enjoy operating in open-ended research environments and want to build entirely new experimental systems from scratch. The ideal profile is someone who combines strong software engineering fundamentals with a genuine curiosity about how models reason, learn, and represent concepts internally.

What they're looking for:

• Strong software engineering fundamentals with experience building complex technical systems
• Experience working on experimental ML systems, research tooling, or model-adjacent infrastructure
• Interest in interpretability, alignment, reinforcement learning, or mechanistic AI research
• Comfort working close to model internals and building custom tooling from scratch
• Research experience is preferred (PhD is a bonus but not required)
• Strong ownership mentality and desire to work in a highly collaborative, research-driven environment

Important details:

• Fully onsite in San Francisco
• $200K-$300K base + equity

Apply to learn more about the team, research, and mission.