Obiguard’s inspection pipeline sits in the critical path between customer applications and their LLMs — every prompt and response passes through it in single-digit milliseconds. As a Software Engineer, you’ll build and harden this pipeline: the detectors, the policy engine, and the audit ledger that enterprise security teams depend on in production.
What you'll do
Design, build, and ship detectors for the inspection pipeline — covering PII, prompt injection, toxicity, and framework-specific policy rules.
Improve p50/p99 latency on the synchronous, in-path inspection layer without compromising detection accuracy.
Own services end-to-end: design, implementation, testing, deployment, and on-call.
Work directly with customers during pilots to debug production incidents and edge cases in real deployments.
Collaborate with the AI Safety Research team to turn new attack research into shipped detectors.
Contribute to the audit ledger and framework-mapping systems that back our compliance guarantees.
What we're looking for
3+ years building production backend systems, ideally in Go, Python, or Rust.
Comfortable working in the critical path of a low-latency system — you think in milliseconds, not seconds.
Experience with distributed systems, observability, and debugging production incidents under pressure.
Solid understanding of security fundamentals; prior application security or infra security experience is a strong plus.
Bias toward shipping — comfortable owning a feature from design doc to production rollout.
Nice to have
Experience building or operating LLM-adjacent infrastructure (RAG pipelines, agent frameworks, model gateways).
Familiarity with compliance frameworks (SOC 2, ISO 27001, ISO 42001, NIST AI RMF).
Open-source contributions to security or infra tooling.
Ready to apply?
Send us your CV. We read every one.
Email [email protected] with your CV and the role you're applying for. We respond to every applicant.