About Me

Building AI systems that actually ship

I got hooked on AI the day I realized you could teach a machine to understand a sustainability question and answer it instantly from a MongoDB database — no analyst required. That was my ESG chatbot, and it's now serving 10+ organizations.

What I got obsessed with since is the harder half: keeping an agent honest once real money and real reporting depend on it. That means the model routes and explains but never computes a regulated figure, and authorisation lives in what the agent can even see rather than in a prompt. Currently at Planet Sustech, where that approach cut manual ESG analysis time by roughly 70%. Every system I ship is one I'd stake my name on.

auto_awesomeWhat I Do

Expertise & Services

psychology

Production LLM Agents

Tool design, role-scoped tool binding, and grounding every figure server-side so the model routes and explains but never invents a number. One agent with 21 typed tools, not a multi-agent system — because a router is another model call with less context.

LangGraphLangChainAWS BedrockRAGAgent EvaluationTool Design
dns

Backend Systems Underneath

The engineering that makes an agent usable: data modelling, aggregation pipelines, and indexing. Took a dashboard query from over three seconds to under one by denormalising at write time — removing 14 $lookup and 8 $unwind stages, cutting documents scanned from ~38K to ~3.2K — and fixing the N+1 beneath it.

MongoDBNode.jsExpress.jsPostgreSQLRedisIndexing
web

Shipping End to End

I own the whole path from server-side APIs to the React/Next.js surface, so an agent reaches real users instead of stopping at a notebook. Shipped production apps with real users from scratch.

Next.jsReactTypeScriptTailwindCSSSupabase
cloud

Cloud & DevOps

AWS-native deployments using Lambda, Cognito, SES, SNS, and EC2. Containerised services with Docker and automated CI/CD pipelines via GitHub Actions.

AWS LambdaEC2DockerCI/CDGitHub ActionsServerless
timelineMy Journey

Career Timeline

school
2019

Started B.Tech

Began Computer Science degree at IPS College

work
2023

First Job

Joined Antino Labs as Associate Software Developer

psychology
2024

Software Developer

Moved to Planet Sustech, focusing on agentic systems

rocket_launch
2026+

Building the Future

Creating production AI systems with multi-agent workflows

codeTech Stack

Skills & Technologies

AI/ML

LangGraphLangChainAWS BedrockRAGAgent EvaluationTool DesignPrompt EngineeringOpenAI GPT-4oNLPSemantic Search

Languages

JavaScriptTypeScriptPythonSQL

Backend & APIs

Node.jsExpress.jsNestJSFastAPIServerless FrameworkREST APIsGraphQL

Frontend

Next.jsReact.jsTailwindCSSServer-Sent EventsZustand

Databases

MongoDBPostgreSQLMySQLRedisSupabasePinecone

Cloud & DevOps

AWS (Lambda, EC2, ECS, Fargate, SES, SNS)DockerCI/CDGitHub Actions

Tools & Libraries

Socket.IOJestPytestGitSwaggerAgora SDKStripeFirebase
starBeyond Code

The Human Side

Fun Facts

☕
Coffee Addict
4+ cups daily
🏏
Cricket & Chess
Sports enthusiast
🎧
Music Lover
Lo-fi & Jazz
🧠
AI Tinkerer
Always experimenting

Daily Drivers

code
VS Code
folder
GitHub
widgets
Docker
send
Postman
description
Notion
terminal
Terminal
psychology
LLM Agents
edit_note
Claude