We're hiring senior software security engineers for a 3-month remote contract building security tasks that train and test frontier AI coding models.

What you'll do

• Take real vulnerabilities (CVEs, security advisories, bug reports) from open-source projects and reproduce them

• Classify each one (CWE, OWASP Top 10) and write proof-of-vulnerability tests

• Build Docker-based audit-and-patch tasks in Python, JS/TS, Java, C/C++, C#, Go, Rust, Ruby, PHP, Kotlin or Swift

• Harden the tests so AI models can't game them, and score AI-written patches

You need

• 5+ years of professional software engineering

• Expert in at least one of the languages above

• A strong grasp of CWE, CVE and the OWASP Top 10

• Hands-on work with sanitizers, fuzzers or static analysis (CodeQL, Semgrep)

• Git, Linux, Bash and Docker

• Clear technical writing in English

• 40 hours a week, with 4 hours overlapping US Pacific time

Pay and length

• $15 per hour

• Fully remote contract, 3 months, starts next week

How to apply

Click Apply and fill in our 2-minute form. A short interest form, an AI interview, then a technical interview. Experienced leads (6+ years, has run a small team) can apply through the same link and say so.