Overview

AI Agent Reliability Engineer @ Emergent
Focus: AI Agents & Backend Systems
--:--_ // IST
he / him

Social Links

GitHub Contributions8,039 TOTAL CONTRIBUTIONS (2,097 THIS YEAR)

Latest Contributions

About Me

About

I engineer high-performance AI systems and design scalable backends that power resilient, production-grade full-stack architectures. My work focuses on building autonomous agents and choreographing complex multi-agent orchestration workflows to solve hard, distributed systems problems.
01.
I'm an AI and backend engineer. Most of my work sits between a language model and the production system it has to run inside.
02.
Currently an AI Agent Reliability Engineer at Emergent, where I debug and fix the agent runs, deployments, and environments behind apps people build with AI.
03.
At CargoFirst I built Blinkus AI, a trade-intelligence platform where seven domain copilots answer questions over live shipments, trade-policy PDFs, and 620K+ customs records.- DeepSeek tool calling, pgvector RAG, and reasoning streamed to the browser over SSE.
04.
At Adaapt.ai I took a B2B lead-generation product from zero to launch.- Semantic matching hit 85% accuracy and cut manual lead sorting by 90%.
05.
At HCLTech I fine-tuned TinyLlama on Azure ML and made inference 35% faster.

"Shipped AI systems at HCLTech and at early-stage startups, and now keeps AI agents running in production at Emergent."

Engineering Focus

"Led frontend delivery for VIT Advitya Fest, supporting 3,000+ active users without crashes during peak load."

Leadership & Execution

Selected Projects

(Showing 4 of 17)

I love designing and building thoughtful, production-grade applications.

View Project

Infinity UI preview

Infinity UI

37+ animated React components you copy into your project instead of installing.

Live
Open Emergent preview

Open Emergent

A self-hostable AI application builder powered by a 9-agent LangGraph mesh architecture for complex agentic workflows.

Building

View Project

SG Agent Harness preview

SG Agent Harness

A CLI that decides when an AI coding agent's work is actually done, by running real checks instead of trusting the agent.

Playwright
Live

View Project

Mugen UI preview

Mugen UI

A shadcn/ui component registry where every component has spring physics, and some have sound.

Motion
shadcn/ui
Live

Tech Stack

Next.js
PostgreSQL
AI / LLMs
Redis
TypeScript
Go
Docker
Rust

Experience

(5)
  • Debugging failed runs of the coding agents that build and deploy apps for Emergent users.

  • Tracing failures across agent runs, deployments, and environments, then fixing the cause so the same ticket doesn't come back.

  • AI Agents
  • LLM Reliability
  • Observability
  • Python
  • TypeScript
  • Built Blinkus AI, a trade-intelligence platform where seven domain copilots answer questions over live shipment data.

  • Wrote the pgvector RAG over trade-policy PDFs, a customs analytics engine over 620K+ records, and SSE streaming of model reasoning to the UI.

  • Next.js 16
  • Supabase
  • pgvector
  • Google Gemini
  • TypeScript
  • Built AI features and the FastAPI and PostgreSQL backend behind them.

  • Containerized services with Docker and deployed them on AWS.

  • Python
  • FastAPI
  • PostgreSQL
  • Docker
  • AWS
  • Built a fine-tuning pipeline for TinyLlama on Azure Machine Learning.

  • Made inference 35% faster and reached 85% accuracy on domain-specific datasets.

  • Set up CI/CD with Azure DevOps and Docker, cutting packaging and deploy time by 40%.

  • Python
  • Azure ML
  • Docker
  • Azure DevOps
  • Took a B2B lead-generation product from zero to launch on Next.js, Supabase, the OpenAI API, and Twilio.

  • Built semantic matching on vector embeddings that hit 85% accuracy and cut manual lead sorting by 90%.

  • Made the REST APIs safe under concurrent load, with zero downtime during peak traffic.

  • Next.js
  • Supabase
  • OpenAI API
  • Twilio

Education

  • Maintained a strong academic record with an 8.55 / 10.0 GPA.
  • Specialized in Data Structures, Algorithms, Distributed Systems, Systems Design, and Web Technology.
  • Served as the Web Development Lead for Advitya Fest, building a registration portal that supported 3,000+ active users without downtime.
  • Data Structures
  • Algorithms
  • Systems Design
  • React.js
  • Java
  • Python
  • SQL
  • C++
  • Java
  • Physics
  • Mathematics

Achievements & Leadership

(2)
Academic Excellence - Maintained an 8.55/10.0 GPA in Computer Science at VIT, with strong grounding in Data Structures, Algorithms, and Distributed Systems.
Web Development Lead (Advitya Fest) - Directed frontend engineering for a React.js registration portal that supported 3,000+ active users without failures during peak load.