About the role
Breaking agentic AI systems for a living: prompt injection attacks, agent orchestration weaknesses, and failure modes in tool use, training data pipelines and guardrails. Every company shipping agents is exposed to attacks nobody has fully catalogued, and this is among the few firms doing serious public work on it.
US remote, Python with LLM, RAG, PyTorch and LangChain, plus threat modelling. Wide band.
Skills
How to apply
Apply directly on the employer's application page. Your application goes straight to them.
More from Trail of Bits
Jobs like this
-
Staff Enterprise Security Engineer, AI Security
Twilio · US only · US remote (pay varies by location band)
-
Staff Threat & Attack Research Engineer
Anduril Industries · US only · United States; ability to obtain and hold a US Secret clearance
-
DevSecOps Engineer
Virta Health · US only · US remote; not hiring in AK, AR, DE, HI, ME, MS, NM, OK, SD, VT or WI
-
Senior Security Engineer, Incident Response
1Password · Canada only · Canada (remote)
Republished listing
This opportunity was discovered on Trail of Bits's public careers page and is republished here for discovery purposes. Applications are handled by the employer.
If this is your role and you want it off the board, contact us.