HarshJannawar
Remote: Yes (remote or hybrid) | Relocate: Yes (SF Bay Area, Seattle, NYC, Austin, other US hubs)
Technologies: Python, security engineering, agentic AI and MCP security, LLM red-teaming, guardrails, threat modeling, DevSecOps, AWS, Docker, Next.js, Node.js, AI security
Résumé: https://www.harshjannawar.com/resume | github.com/Harshj143 | [email protected]
What I’m proudest of is taking a startup from zero to a live production platform with 20+ paying customers within 3 months, built entirely from the ground up: product development, secure SDLC, full-stack delivery, and the AWS infrastructure to run it daily.
Inline Proxy is my open source MCP security gateway: it enforces allow/block/redact/approve policy on every agent tool call and blocks a full prompt-injection exfiltration chain without ever needing to detect the injection itself.
On the offensive side, I designed and ran a prompt injection payload against a Claude Opus 4.8 email triage agent in a controlled lab, achieving successful injection in 8 of 10 attempts while the agent still completed its legitimate task in all 10. I also cleared all 3 escalating levels of Trajectory Labs’ Terminal red team assessment against a hardened PARALLAX agent, including a ticket authority spoof that triggered a privileged password reset. Trajectory Labs’ lab work has found vulnerabilities across Opus 5.5, Fable, Mythos, Gemini, and Argon.
MS in Cybersecurity Engineering (UW), thesis became AegisLLM, an OWASP LLM Top 10 red team suite using a genetic algorithm to evolve attacks: github.com/Harshj143/AegisLLM. Top 1% TryHackMe, CTF win at UWB GreyHats.
Looking for security engineering or AI security roles at companies putting agents in production.
Remote: Yes (remote or hybrid)
Willing to relocate: Yes (SF Bay Area, Seattle, NYC, Austin and other US hubs)
Technologies: Python, security engineering, agentic AI and MCP security, LLM red-teaming, guardrails, threat modeling, DevSecOps, AWS, Docker, Next.js, Node.js
Résumé/CV: https://www.harshjannawar.com/resume | github.com/Harshj143
Email: [email protected]
AI security engineer. I took a product from zero to production and 40+ paying clients in a month, and I build the controls that stop agents from doing damage once they have real tool access.
What I'm proudest of is an open-source MCP security gateway (Inline Proxy). It sits between an agent and its tools and enforces policy on every call: allow, block, redact, require approval. The case it exists for: a poisoned GitHub issue tells the agent to publish the repo's .env as a public gist. The gateway never has to detect the injection. It only tracks that untrusted content was read, so the exfiltration write gets blocked. Default-deny packs cover 194 tools across GitHub, Jira and Slack, ~0.1ms p99 on the enforcement path, fail-closed. [Inline Proxy repo link]
My master's thesis (UW 2026, ProQuest 32738510) became AegisLLM, an LLM red-team suite mapped to the OWASP LLM Top 10 that uses a genetic algorithm to evolve new attacks. github.com/Harshj143/AegisLLM
MS in Cybersecurity Engineering (UW), top 1% on TryHackMe, CTF win at UWB GreyHats. Looking for security engineering or AI security roles at companies putting agents in production. Happy to walk through the design of any of this.
Remote: Yes (remote or hybrid)
Willing to relocate: Yes (SF Bay Area, Seattle, NYC, Austin, other US Hubs)
Technologies: Python, DevSecOps, security engineering, security automation, threat modeling, SAST/DAST/SCA, AWS, Docker, Next.js, Node.js, OWASP LLM Top 10, AI/LLM security, CI/CD security, PostgreSQL, Hugging Face, LangChain, LLM APIs (OpenAI/Anthropic), FastAPI.
Résumé/CV: https://www.harshjannawar.com/resume | https://github.com/Harshj143
Email: [email protected]
Most AI agents ship with tool access and nobody watching what they do with it. I build the enforcement layer that catches it before it becomes a breach.
Inline Proxy: open-source MCP security gateway enforcing per-tool policy (allow, block, redact, and more) on every agent tool call. Taint tracking stops a poisoned GitHub issue from publishing a repo's .env as a public gist. 194 tools covered, ~0.1ms p99, fail-closed by default.
AegisLLM: published as my master's thesis (UW 2026, ProQuest 32738510). Automated LLM red-teaming mapped to the OWASP LLM Top 10, with a genetic-algorithm engine that evolves new attack variants.
MS in Cybersecurity Engineering from UW. Top 1% on TryHackMe, with multiple CTF wins. Looking for security engineering or AI security roles at companies putting agents in production.