I build production AI agents and the data systems behind them.
I design multi-agent GenAI systems: orchestrator agents, text-to-SQL with guardrails, and LLM-ready data layers, shipped on Kubernetes with tracing and evals.
- Now
- Development Engineer 3 at Comcast
- Experience
- 6+ years in software, the last 1.5 years on production GenAI
- Before GenAI
- Ruby on Rails backends at Comcast and TCS
- Education
- M.Tech, Computer Science (VIT Chennai), with a published deep learning paper
- Based in
- Chennai, India
Featured work
Patterns from production systems, written up as case studies. Details are generalized; no confidential data.
Orchestrator agent with specialist agents
An executive reporting agent that understands the question and delegates to specialist agents for data, visualization, business context and email.
Config-driven use-case engine
Use cases stored as versioned database records. New use cases ship with no code changes, and deterministic steps skip the LLM.
Text-to-SQL with guardrails
Read-only queries, row limits, schema-grounded columns and an entity map from business terms to columns.
Preparing a database so an AI agent can query it
The agent writes SQL from plain-English questions. I prepared the database so its queries are correct and fast: column descriptions it can read, clean views, a pre-computed table and search indexes.
More things I've built
Change-review agent
Reviews change requests before approval for risk, schedule conflicts and maintenance-window compliance, with alerts for high-risk changes.
Visualization agent
A standalone agent that other agents call as a tool to pick, validate and render the right chart for their data.
Dashboards from a question
The agent returns a validated JSON dashboard spec, rendered from reusable components with no new frontend code.
Asset intelligence agent
A supervisor agent over infrastructure asset data, with intent routing, GraphRAG dependency analysis and episodic memory.
n8n to LangGraph migration
Moving agent workflows to LangGraph services with streaming FastAPI, checkpointing, tracing, Kubernetes and GitHub Actions. One system is done, the second is in progress.
Shared building blocks
A DeepEval answer benchmark, document extraction, pgvector RAG ingestion, a React and FastAPI chat interface, and a report-deck generator.
Skills
What I work with.
Experience
6+ years: from Rails backends to production GenAI.
Development Engineer 3 · Comcast
Multi-agent systems, text-to-SQL, n8n to LangGraph migration, PostgreSQL data layer, generative dashboards.
Development Engineer 2 · Comcast
Ruby on Rails and Vue.js services that turn 20+ data sources into dashboards and alerts; CI/CD, upgrades and observability.
System Engineer · Tata Consultancy Services
Rails and GraphQL backend features for a platform used across thousands of stores; Azure CI/CD automation.
M.Tech, Computer Science · VIT Chennai
Published research on malware classification with lightweight CNNs (Springer). Read the paper →