Vertical Relevance is partnering with a leading financial services organization to build an innovative AI-driven security capability that transforms how application vulnerabilities are identified, validated, and prioritized. The solution uses agentic AI, static analysis, reachability analysis, and isolated execution environments to generate executable proof of exploitability and publish validated findings through existing security workflows.
What You Will Do
- 7+ years Design and build an agentic vulnerability-discovery harness from end to end.
- Architect the agent runtime, orchestration layer, and parallel vulnerability-hunting agents.
- Develop adversarial validation workflows that verify candidate findings before publication.
- Build cross-repository indexing, dependency mapping, and source-to-sink reachability analysis capabilities.
- Establish isolated execution environments for exploit validation.
- Validate the platform against labeled vulnerability datasets and measure detection quality.
- Publish validated findings into Azure DevOps using SARIF.
- Create runbooks, hardening guides, and production-readiness documentation.
Required Qualifications
- Production experience building LLM agent or multi-agent systems.
- Experience with agent orchestration, tool integration, evaluation loops, and non-deterministic model behavior.
- Strong software architecture and engineering experience.
Preferred Qualifications
- Application security, vulnerability discovery, and exploit development.
- CodeQL, Semgrep, Joern, taint analysis, and data-flow analysis.
- Tree-sitter, SCIP, or LSP-based indexing.
- Firecracker, gVisor, or similar isolated execution platforms.
- SARIF and Azure DevOps integration experience.
Technology Environment
LLM Agent Frameworks: LangGraph, Bedrock AgentCore, MCP-based tooling
Security Analysis: CodeQL, Semgrep, Joern, SAST, SCA
Isolation Technologies: Firecracker, gVisor, Containers
DevOps & Security Operations: Azure DevOps, SARIF, CI/CD Security Pipelines
