Member of Technical Staff - Research, AI Safety

Help Define the Future of AI Alignment

As a Research MTS, you will spearhead the creation of high-stakes benchmarks designed to detect sophisticated flaws in advanced AI, such as strategic deception and reward hacking. You aren't just writing code; you are architecting the infrastructure that ensures the next generation of agents—the same ones we use for vibe coding—remain safe and reliable. By collaborating with elite foundation model labs, your work will directly influence the global standards for AI behavior.

Technical Landscape & Vibe Culture

This role is a perfect match for researchers who leverage AI tools to amplify their technical output. You will work primarily in a Python and Django environment to construct complex evaluation scenarios, a stack that is highly compatible with modern AI coding assistants.

• Rapid Iteration: Use AI-assisted development to quickly prototype and deploy new safety tests and research environments.
• Modern Infrastructure: Manage scalable research systems on AWS using CDK for Infrastructure as Code.
• Agentic Research: Design and analyze multi-agent systems to observe how frontier models interact in real-world simulations.

Why Join Vals AI?

This is a high-intensity, talent-dense environment born out of Stanford research and backed by a $40M Series A. You will have the autonomy to tackle massive safety challenges without waiting for consensus, benefiting from a culture that values extreme ownership and solution-oriented thinking. Beyond the impact, you’ll enjoy a San Francisco-based hub with premium perks, including housing stipends, provided meals, and significant equity ownership.