Software Engineer - AI Security Product
Obsidian Security · Palo Alto, California, USA
About the role
Software engineer on Obsidian Security's AI security product, defining how the platform spots and reduces risk as tools such as Copilot, ChatGPT Enterprise, Gemini, Claude and agentic frameworks connect to corporate SaaS. The engineer builds expertise in how these platforms authenticate and what scopes they request, researches prompt injection, agent misuse, RAG poisoning and over-permissioned integrations, owns threat and data models, prototypes and ships AI agent security features, and improves data pipelines.
Obsidian wants one to three years shipping production backend systems in Python or Go, cloud, container and CI/CD experience, hands-on LLM APIs and agent patterns, and a modern data stack such as dbt with ClickHouse or Snowflake. The role is based in Palo Alto.