About Me
AI engineer with 5+ years of experience building production-grade web, mobile, and AI agent systems. I built Mail Pilot, an AI-native email client with CopilotKit and tool-based AI architecture; Quartermaster, a conversational AI agent with 26 tools built on LangGraph/deepagents; and Aegis, an AI debugging system that translates natural language to SQL and CloudWatch queries. Also scaled WatchWithMe to 7,000+ users. Each system ships with full technical documentation.
I care about measurable outcomes — the same engineering discipline goes into my AI agent systems. At Fluid.travel (PickYourTrail), I revamped the hotel booking flow driving a 150% increase in bookings. At Klenty, I rebuilt the mail tracking system reducing latency from 15 minutes to 5 seconds. At Codedamn, I cut infrastructure costs by 55% through container allocation optimization.
My core strengths are AI agent architecture, LLM orchestration, real-time systems, and observability. I was an invited speaker at Chennai Web Conf (Dec 2025) where I presented "Artemis: Personal AI OS", covering AI-native applications and personal AI system design.
Experience
Frontend Developer (SDE II)
- Owned end-to-end mobile app development using React Native
- Migrated React Native 0.76 → 0.82, improving performance and stability
- Revamped hotel search & booking flow, resulting in 150% increase in bookings
- Built cancellation & reschedule policy module
- Developed Reports module for operational insights
- Automated GST invoice retrieval from Indigo flights (100–400 daily bookings)
- Built production AI agent systems (Aegis, Mail Pilot, Quartermaster) alongside mobile development — LangGraph, LLM orchestration, and human-in-the-loop architectures
Fullstack Developer (Consultant)
- Built offline sync functionality for mobile app
- Implemented multi-session sync across devices for MSMEs
- Developed SMS parsing for automated transaction recording
Full Stack Developer
- Implemented Zipkin distributed tracing for enhanced performance measurement
- Built real-time mail tracking system, reducing latency from 15 minutes to 5 seconds
- Designed MongoDB request attribution system for precise per-request observability
- Built slow query detection pipeline for infrastructure optimization
- Enabled precise debugging of performance bottlenecks across the stack
Senior IDE Engineer
- Built DSA playground with custom code execution stack
- Developed real-time collaborative coding editor
- Implemented autocomplete & IntelliSense features
- Reduced infrastructure cost by 55% through container allocation optimization
Member of Technical Staff
- Built advanced pivot table analytics system for cross-module data aggregation
- Implemented conditional formatting & script columns
- Enabled spreadsheet-like computations inside product
Skills
Languages
Frontend
Backend
Databases
AI & ML
Cloud & DevOps
Core Strengths
Flagship Projects
Production-grade systems built from the ground up, with full technical documentation
Mail Pilot
FlagshipAI-native email client
An AI-native email client where you control everything through natural language. The AI assistant navigates views, filters threads, composes replies, searches your inbox, and manages labels, while a human-in-the-loop approval system guards every action.
Key Features
Tech Stack
Explore the Docs
Quartermaster
FlagshipConversational kirana-store agent
A conversational AI agent that lets a kirana store owner run their entire business through Telegram in plain language. Billing, inventory, khata, GST invoices, analytics, and reminders — all through chat, no web app needed.
Key Features
Tech Stack
Explore the Docs
Aegis
FlagshipAI-powered production debugging system
A dual-mode AI framework that lets you query and debug production systems using natural language. Connect databases, logs, and infrastructure — ask questions in plain English, get answers with code, charts, and actionable insights.
Key Features
Tech Stack
Explore the Docs
Stardust.app
FlagshipContainer provisioning engine on AWS Spot Instances
A complete container deployment platform that takes you from a GitHub repository to a running container on AWS — with real-time monitoring, SSH access, and disaster recovery. Built on Spot Instances with on-demand fallback.
Key Features
Tech Stack
Explore the Docs
Projects
WatchWithMe
Real-time watch party platform (Web & Mobile). Scaled to 7,000+ users with paying customers. Built synchronized playback using WebSockets, integrated voice & video chat, and solved hard latency, drift & onboarding challenges.
7,000+ users • Paying customers
Job Genie
AI-powered toolkit designed to help job seekers apply smarter, not harder. Includes tools for generating personalized cover letters, email builder, and resume optimization for senior roles.
Blog
I write about building AI-powered tools, real-time systems, observability, and full-stack engineering.
The Case Against 'Don't Reinvent the Wheel'
Why building complex systems from scratch is the best way to truly understand architecture.
Building AI Agents That Actually Work
Tool composition, closed feedback loops, and making AI an engineer you empower.
Real AI Coding Workflows: Beyond Figma Dumps
How to structure AI coding agents with micro-component prompting and proper guidelines.
Get in Touch
I'm always interested in hearing about interesting projects and opportunities. Feel free to reach out!