Interpretability Researcher
LB322
Posted: 10/11/2026
- $200,000 - $600,000
- California
- Permanent
Interpretability Researcher - Mechanistic Interpretability
New York, NY / San Francisco, CA (On-site)
I'm working with a well-funded frontier AI research company exploring how AI agents can automate and scale mechanistic interpretability research.
Company highlights:
- Funding: Series A | $60M raised
- Valuation: $725M
- Investors: Sequoia Capital, Menlo Ventures, BCV and SV Angel
- Locations: New York City & San Francisco
The team works directly with frontier AI labs, developing reinforcement learning environments, evaluation systems and research infrastructure for autonomous AI agents.
The role:
- Investigate model internals, neural circuits and learned representations
- Develop AI agents to automate mechanistic interpretability experiments
- Build research tools, evaluation frameworks and experimental pipelines
- Explore training tasks and feedback mechanisms that improve AI-assisted research
- Collaborate with researchers across interpretability, reinforcement learning and AI alignment
What they're looking for:
- L5 - L7+ hires
- Research experience in mechanistic interpretability, model internals or representation learning
- Hands-on experience with activation analysis, sparse autoencoders, circuit discovery or causal interventions
- Strong Python/PyTorch skills and practical research engineering capabilities
- Experience with AI agents, automated experimentation or research tooling is a strong advantage
- Original research, publications or meaningful open-source contributions
Why consider it?
An opportunity to join a heavily backed AI research team tackling a fundamental challenge: can frontier AI models accelerate our understanding of how neural networks work internally?
Location: NYC or SF | On-site
Interested? Apply directly or message me for a confidential discussion.
Zee Uddin
Researcher