Security Researcher

Eden Prescott; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Eden Prescott is partnering with an AI research lab building foundation models for cybersecurity. The team is looking for a Security Researcher who can bring deep security expertise to model development, helping AI systems discover vulnerabilities, reason about complex software, and produce findings that stand up to technical scrutiny.

This role sits at the intersection of vulnerability research, reverse engineering, and applied AI.

Qualifications

• Evidence of original security research: vulnerability discoveries, technical write-ups, research tools, open-source contributions, or substantial private work you can discuss.

Responsibilities

• Conduct original vulnerability research across software, systems, and open-source projects in authorized research environments.
• Build reproducible demonstrations that distinguish exploitable vulnerabilities from false positives.
• Partner with ML researchers to translate security research workflows into training tasks and feedback signals.
• Design realistic evaluations that test models’ ability to understand code, identify vulnerabilities, and validate findings.
• Develop tools and environments for automated analysis, fuzzing, debugging, and experimentation.
• Investigate where models fail and help improve their reasoning, tool use, and technical accuracy.
• Support responsible disclosure and contribute to research publications where appropriate.

Skills

Vulnerability research

Reverse engineering

Tooling development

Independent work

Collaboration

Tools

Fuzzing tools

Symbolic execution tools

LLMs in research workflows

Automated vulnerability discovery

Job description

Eden Prescott is partnering with an AI research lab building foundation models for cybersecurity.

The team is looking for a Security Researcher who can bring deep security expertise to model development, helping AI systems discover vulnerabilities, reason about complex software, and produce findings that stand up to technical scrutiny.

This role sits at the intersection of vulnerability research, reverse engineering, and applied AI. You’ll work alongside researchers and engineers to turn security expertise into training environments, evaluation methods, and measurable improvements in model capability.

What you’ll do

• Conduct original vulnerability research across software, systems, and open-source projects in authorized research environments.
• Build reproducible demonstrations that distinguish exploitable vulnerabilities from false positives.
• Partner with ML researchers to translate security research workflows into training tasks and feedback signals.
• Design realistic evaluations that test models’ ability to understand code, identify vulnerabilities, and validate findings.
• Develop tools and environments for automated analysis, fuzzing, debugging, and experimentation.
• Investigate where models fail and help improve their reasoning, tool use, and technical accuracy.
• Support responsible disclosure and contribute to research publications where appropriate.

What we’re looking for

• Evidence of original security research: vulnerability discoveries, technical write-ups, research tools, open-source contributions, or substantial private work you can discuss.
• Strong programming skills and the ability to build tooling that supports your research.
• A rigorous approach to validating findings and explaining how a vulnerability works.
• Curiosity about foundation models and how they can improve security research.
• The ability to work independently on ambiguous problems and collaborate closely with people outside your specialty.

You don’t need to have trained a foundation model. You do need strong security research instincts and an interest in working closely with the people who do.

Particularly relevant experience

• Fuzzing, symbolic execution, static or dynamic analysis.
• Research involving memory corruption, complex application logic, or multi-step exploit chains.
• Building automated vulnerability discovery or validation systems.
• Using LLMs or agents in technical research workflows.
• Creating benchmarks, research datasets, or reproducible testing environments.

Why this role

• Help shape models built specifically for cybersecurity.
• Work on open research problems alongside security and post-training specialists.
• Influence how the team defines, measures, and improves model capabilities.
• Build tools that could make rigorous security research more scalable.