Vals Ai
Member of Technical Staff - Research, AI Safety
San Francisco. Posted 25 days ago
The role
You will design and build novel benchmarks and environments to assess misaligned behavior in frontier models and agents. You will conduct research, collaborate with foundation model labs, and publish findings.
What you will do
- Design and develop novel, high-impact benchmarks and environments that assess real-world misbehavior in frontier models and agents
- Conduct research to ensure benchmarks are valid, reliable, and meaningful
- Collaborate with foundation model labs and enterprises to understand evolving safety evaluation needs
- Analyze model behavior across benchmarks and articulate results
- Publish research findings and contribute to the broader safety and evaluation research communities
- Stay current with the latest developments in LLM capabilities and alignment evaluation methodologies
Required qualifications
- Learning velocity
- Ownership
- Intensity
- Solution-oriented mindset
Nice to have
- Substantial, hands-on work in safety evaluations, monitorability, scheming, deception, or multi-agent safety
- PhD focused on AI safety
- Published safety research
- Deep ML/AI background
- Excellent coding skills in Python
- Ability to clearly communicate complex research ideas
- Mechanistic interpretability experience
- Contributions to open-source safety or evaluation tools and datasets
Benefits
- Highly competitive salary
- Meaningful ownership
- Relocation and transportation support
- Health/dental insurance coverage
- Lunch and dinner provided
- Free snacks/coffee/drinks
- Unlimited PTO
- $1,500 housing stipend (within one-mile radius)
About the company
The core methodology behind this platform comes from NLP evaluation research conducted at Stanford. The team has prior work experience at NVIDIA, Meta, Microsoft, Palantir and HRT.