Security Architect - AI-Powered Security Engineering

Security Architect - AI-Powered Security Engineering

About the Team

The security team at Meesho is like the Avengers to Meesho's S.H.I.E.L.D. When 5% of Indian households shop with us, it’s important to build resilient systems to manage millions of orders every day. We’ve done this, with zero downtime! We value speed over perfection, and see failures as opportunities to become better. By 2030, Meesho will be the world’s most trusted and secure digital ecosystem, operating on an AI-native, Zero-Trust, self-healing architecture. As we shift toward an era where autonomous agents generate and deploy code, we want to ensure our security scales seamlessly with engineering velocity. We place special emphasis on the continuous growth of each team member, fostering a strong 'Founder’s Mindset' that helps us move fast.

About the Role

We are looking for a highly technical, hands-on Security Architect (Individual Contributor) to guide and lead our AI-Powered Security Engineering charter. In this role, you will bridge the gap between conventional DevSecOps, infrastructure security, and the rapidly evolving landscape of AI. You will optimize our AI-driven Software Development Lifecycle (SDLC) by designing secure-by-default guardrails for both human developers and AI coding agents. Your mission is to ensure near-zero vulnerabilities from application and data security failures reach production. A significant portion of your focus will involve advanced Red Teaming, utilizing LLMs to improve Meesho's security posture, and architecting defenses for our autonomous AI systems against prompt injection, data leakage, and jailbreaking attacks.

What You Will Do: Lead AI Security Architecture: Drive the technical roadmap and high-level design to continuously prevent, detect, and autonomously contain threats across our systems and AI agents.

Hands-on Prototyping & POCs: Build proof-of-concepts for autonomous agent harnesses that automatically detect, analyze, and mitigate security vulnerabilities in real-time.

AI SDLC Optimization: Architect non-bypassable security guardrails across the SDLC for both developers and AI coding agents to ensure production systems are secured by default.

Red Teaming & Adversarial Validation: Proactively execute adversarial testing by simulating real attack scenarios across APIs, AI systems, and infrastructure to validate defense mechanisms.

Protect AI Systems: Design deep technical defenses to protect all AI agents and experimental features from prompt injection, system misuse, manipulation, and unauthorized actions.