About the Role
Corridor is redefining product security for the AI era by building the security layer on AI-generated code. This role focuses on product AI problems, including improving evals, shipping improvements, and owning work end-to-end.
Responsibilities
- Own and improve our evals.
- Ship product changes that enhance AI systems, including prompts, retrieval, agent scaffolding, and model/tool integrations.
- Own work end-to-end, from prototype to production.
- Collaborate with AI and product engineers to integrate new AI capabilities into the platform.
Requirements
- Experience shipping production code and owning features end-to-end.
- A track record of building on LLMs and measurably improving AI systems.
- Hands-on experience with LLM evaluation.
- 3+ years of experience in a software engineering or ML role (ideally 5+).
- Self-directed and product-minded with the ability to own ambiguous projects.
Skills
- LLM evaluation
- Building on LLMs
- Improving AI systems
- Infrastructure or DevOps experience (CI/CD, cloud, observability)
- Model fine-tuning
- Code generation
- Security
Location
- Hybrid
Work Type
- Hybrid
- Full-time
Experience Level
- 3+ years of experience in a software engineering or ML role
- Ideally 5+ years of experience
About the Company
- Corridor is redefining product security for the AI era.
- Our team works at the intersection of AI and cybersecurity.
- Our leadership has extensive experience in security at major companies, government policy, and AI research.
- Co-founder and CEO Jack Cable is a top-ranked bug bounty hunter who previously led Secure by Design at CISA.
- Co-founder and CTO Ashwin Ramaswami built large-scale systems at Skiff, Caldera, and Nooks, and published research on AI and open foundation models at Stanford.
- CPO Alex Stamos is the former CISO of Facebook, Yahoo, and SentinelOne and a current lecturer at Stanford University.
- Notable angels and advisors include Russell Kaplan (Cognition), Thomas Wolf (Hugging Face), Jonathan Frankle (Databricks), and Mike Krieger (Anthropic).
