About this role
Continuum is autonomous security at machine speed: discover, validate, prioritize, and remediate code vulnerabilities. You will own the evaluation and release systems that determine what ships and when. What we're looking for Strong engineering fundamentals with experience in ML or evaluation infrastructure Hands-on experience with experiment design, A/B testing, or online evaluation Comfort at the boundary of research and production Experience with agent systems, LLM evaluation, or ML pipelines is a plus Key job responsibilities Production evaluation systems Build automated pipelines that evaluate changes to Continuum's models, prompts, tools, and agent architecture. Provision realistic environments, run deployed agents, grade results, and compare candidates against baselines. Design for asynchronous, nondeterministic agent behavior. Create reusable workflows so new capabilities follow a standard evaluation and release path. Online experiments and release decisions Design Continuum's A/B testing and online evaluation system. Support shadow evaluation, canary releases, and controlled experiments. Build treatment assignment, exposure logging, guardrail metrics, stopping criteria, and rollback. Define metrics for quality, safety, latency, cost, and reliability. Build dashboards that surface regressions and produce readouts that explain what changed and what to monitor post-launch. Shipping agent capabilities Turn prototypes into production features alongside scientists and security engineers. Package changes to models, prompts, tools, planning, memory, and agent coordination behind stable interfaces. Own science-side integration end to end: APIs, data contracts, versioning, telemetry, release tests, and production requirements for tenancy, durability, scaling, and failure recovery.