SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Scale is seeking a Senior AI Product Manager to build and own Scale's Cybersecurity portfolio—a new product line focused on data, environments, and evaluations that frontier AI labs use to train and measure security capabilities in their models.
This is a foundational build role where you will define strategy and standards for a product line that does not yet exist. Security represents one of the hardest problems in agentic AI: agents capable of finding vulnerabilities, reproducing them, and generating patches without breaking systems perform work that typically requires skilled humans days to complete. Measuring this capability honestly requires reproducible execution environments at scale and practitioners with real security expertise.
Key responsibilities include:
- Owning the complete roadmap and strategy for the Cybersecurity portfolio across training data, RL environments, agentic task suites, and evaluation products
- Defining the capability map for training and measurement: vulnerability discovery, proof-of-concept reproduction, patch generation, secure code review, supply-chain analysis, malware analysis, detection engineering, and incident triage
- Making strategic decisions on where Scale competes across the offense-defense spectrum
- Partnering with ML researchers and security practitioners on task specifications, grader design, and verifiable rewards
- Driving the infrastructure roadmap including reproducible vulnerability images, fuzzing toolchains, sandboxed execution, and automated verification
- Owning responsible-development posture including containment, coordinated disclosure, and sensitive artifact handling
- Establishing governance for data quality, contamination prevention, license hygiene, and reproducibility
- Recruiting and stewarding a contributor network of vulnerability researchers, exploit developers, malware analysts, and detection engineers
- Managing external partnerships with open-source benchmarks, academic security groups, and enterprise partners
- Working directly with frontier labs and enterprise customers to understand model failures and translate insights into roadmap
Ideal candidates bring real cybersecurity work experience (vulnerability research, fuzzing, exploit development, malware analysis, red teaming, detection engineering, or incident response), 5+ years in product management or technical roles, familiarity with the AI-for-security evaluation landscape, software engineering depth, understanding of model post-training and evaluation, strong stakeholder management skills, and sound judgment on dual-use considerations.
