jobs
Machine Learning Infrastructure Engineer, Safeguards Research
Anthropic’s Safeguards team builds systems that detect and mitigate misuse of its AI models. Some of this work relies on lightweight detection methods trained on model internals, so researchers can…
In inglese
- Company
- Anthropic
- Location
- San Francisco, CA / New York City, NY
- Status
- Closed
- Posted
- 2026-11-12 03:34 UTC
Anthropic's Safeguards team builds systems that detect misuse of its models; this engineer owns the infrastructure behind that research. The work covers data pipelines, training, evaluation and scoring workflows, libraries and command-line tools for researchers, correctness checks, and moving experimental workflows to production-grade jobs while improving inference throughput, cost and reliability. It asks for strong Python and software engineering, production experience with data-intensive or distributed systems, and tooling used by other researchers. It sits between research and production, in San Francisco or New York, with at least 25% office time.
Original job posting