I build secure AI agents and multi-agent systems.
The proof is below: production AI agents with real numbers, interactive project views, and an AI assistant that answers from my resume. Scroll — or just ask.
Where I work.
Projects.
Production agents, the security around them, and the systems they run on. Click any project to expand. Research items are marked under review.
In pilot with the operations team (~50 users), with rollout to 12,000+ AWS sellers underway over the next ~6 months. Led the technical design end to end and was part of the team that built it. Integrated LLMs through Amazon Bedrock with dual paths — MCP Server and AgentCore Runtime (Strands) — plus Text2SQL and vector search. Security owned end to end: Bedrock Guardrails, prompt-injection filters, and memory-poisoning safeguards, with formal application-security approval before launch.
Explore, hands-on.
The same interactive views the chat renders inline as MCP Apps — also live here, hands-on.
- Led the technical design of BOB end to end and was part of the team that built it — a production AI agent for AWS sellers (compensation and policy questions in natural language), in pilot with the operations team (~50 users), expanding to 12,000+ sellers over ~6 months.
- Delivered the first prototype in 6 weeks — 40% ahead of a 10-week estimate — then iterated toward rollout.
- Integrated LLMs through Amazon Bedrock with dual paths — MCP Server and AgentCore Runtime (Strands) — plus Text2SQL and vector search.
- Owned security, evaluation, and observability: Bedrock Guardrails, prompt-injection filters, and memory-poisoning safeguards, with formal application-security approval before launch.
- Built a 10-agent SDLC pipeline on Kiro using AgentCore Runtime and Strands, distributed company-wide for reuse across Amazon teams.
- Led application security reviews and penetration testing for the Varicent SaaS migration for 12,000+ users.
Secure AI agents
you can audit.
I build production AI agents at AWS in Seattle — there since December 2016, with 10+ years taking complex products from concept to production. My home domain is Incentive Compensation Management: the systems that define, calculate, and reconcile how sellers get paid.
Lately my work is applied AI inside that domain. I led the technical design of BOB — an AI agent that answers sellers' compensation and policy questions — end to end, and was part of the team that built it. Separately, I built a 10-agent SDLC pipeline for agentic app development, now reused across Amazon teams. I sit on application security reviews, and I bring the same discipline to agents: threat-model the tool surface the way you'd threat-model a service.
An invention disclosure from this work is under review at Amazon — in plain terms, it's about letting a multi-agent plan adapt its own task dependencies as the work runs, instead of following a fixed graph.
What draws me is building AI agents that are more secure and more capable — and this site is how I work in practice: read-only tools, database-level role separation, and honest labels on what's shipped versus what's under review.
- 2016 — nowAWS, Seattle — Application Development Engineer, ICM. Production AI agents, agent security, sales-compensation systems.
- 2016Hughes Network Systems — led a Salesforce Service Cloud integration.
- 2015M.S. Computer Science, UNC Charlotte — plus a retail-analytics internship at Tresata.
- 2010 — 2014HCL Technologies — banking features for Commonwealth Bank of Australia (onsite in Sydney); mentored 15 junior developers.
- 2010B.E., Anna University, Chennai.