Technical Lead — Full Stack · Jalandhar, India
Leading teams across the stack,shipping production systems.
Full-Stack Technical Lead with 9+ years building and shipping production web, mobile, and browser-based applications. Open-source contributor to Next.js (vercel/next.js), builder of agentic AI systems with LangChain and Mastra AI, and author of apps live on the Chrome Web Store and Google Play.
Years Shipping
Users Reached
Next.js Contributor
Apps on Play Store
Experience
Where I've worked
DigiMantra Labs Pvt. Ltd.
2022 – Present
Associate Technical Lead
Lead full-stack architecture and delivery across 2–3 concurrent client projects spanning React/Next.js frontends, Node.js/Nest.js and Python backends, and containerised deployments. Technically lead a team of 6 engineers — conducting architecture and API design reviews, reviewing 6–8 pull requests weekly, and mentoring 10+ junior and mid-level engineers.
Erosteps Pvt. Ltd.
2017 – 2022
Senior Full Stack Web Developer
Built and shipped 8–10 production full-stack web applications, owning both the Vue.js frontend and the Lumen (PHP) and Node.js API layers. Built and maintained CI/CD pipelines using Jenkins and Docker. Shipped 2 cross-platform mobile apps to the Google Play Store using Flutter.
Engineering Philosophy
How I think about building software
01
Performance is a feature
Every millisecond of load time costs real users. I treat Core Web Vitals as hard constraints from day one — not an afterthought before release. A fast product is a respectful product, and the engineering choices that enable speed compound across the entire codebase.
02
Design and engineering are one discipline
The best interfaces emerge when the person writing the code understands why a design decision was made, not just what it looks like. I invest in design-token systems, shared language with designers, and close feedback loops that keep intent intact from Figma to production.
03
Systems thinking over clever solutions
Clever code impresses in a PR review; well-structured systems hold up through two years of team turnover and a product pivot. I optimise for clarity, composability, and constraints that make the wrong choice harder to make than the right one.
Selected Work
Things I've built
Study.IQ — EdTech Platform
Led the React→Next.js migration and payment/conversion overhaul for India's largest EdTech platform serving 1M+ users — driving a 9x web conversion lift (1.16% → 11.3%) and 71% organic traffic growth.
Guardian Services — Security Platform
Architected and built the full platform backend for a security services company — data model, REST API, service-scheduling logic, RBAC, and encryption, owned end-to-end.
Kanban Tab — Chrome Extension
A Chrome Extension replacing the new-tab page with a unified Kanban board syncing Jira, GitHub, Asana, and Linear in real time — built end-to-end and live on the Chrome Web Store.
Projects
Things I've shipped
GitHub PR Reviewer
Agentic AI System
LangChain and Mastra AI agent that ingests a GitHub PR URL, performs structured code review, and posts inline comments via the GitHub API.
- –LangChain and Mastra AI agent that ingests a GitHub PR URL, performs structured code review, and posts inline comments via the GitHub API.
- –Designed the prompt orchestration, tool-calling pipeline, and API integration end-to-end.
Kanban Tab
Chrome Extension — New Tab Kanban Board
Replaces the new-tab page with a unified Kanban board syncing Jira, GitHub, Asana, and Linear in real time. Built end-to-end with Manifest V3 and OAuth 2.0 flows for four integrations.
- –Replaces the new-tab page with a unified Kanban board syncing Jira, GitHub, Asana, and Linear in real time.
- –Built end-to-end: Manifest V3 architecture, OAuth 2.0 flows for four integrations, and the real-time sync layer.
Study.IQ
EdTech Platform — 1M+ Users
Led the React.js to Next.js migration and end-to-end rebuild of the checkout and payment flows — engineering backbone of a 9x web conversion lift.
- –Led the React.js to Next.js migration — porting 30–40 routes, adopting App Router, refactoring to server components, and rebuilding the data-fetching layer for SSR; lifted Lighthouse Performance from ~40 to 75+.
- –Integrated Razorpay as the primary payment gateway with Paytm and CCAvenue fallbacks — building webhook handling, retry logic, and failover on the backend — supporting a payment drop-off reduction from 9% to 4%.
- –Built Continue Purchase Logic and rebuilt PDP, PLP, and Checkout flows end-to-end — the engineering backbone of a conversion initiative that drove a 9x web conversion lift (1.16% → 11.3%) measured by the product team.
Latest Writing
From the blog.
What Is an AI Agent Harness? I Followed Tejas Kumar's "Poor Man's Harness" Build to Find Out.
Tejas Kumar's talk "Harnesses in AI: A Deep Dive" finally made it click: an AI agent harness is the scaffolding around the model, and it's where the real engineering lives. So I followed his "poor man's harness" build step by step — porting it into TypeScript and the Claude API as a browser agent that drives Playwright to navigate Hacker News. The agent loop took twenty minutes. The real work was the tool registry, context compression, a hard step cap, a verify step so the agent can't lie about finishing, and deterministic auth so the model never touches a login. Here's the build, what each piece is for, and why he calls 2026 the Year of Harnesses.
Matt Pocock Is Right: AI Makes Software Fundamentals Matter More, Not Less.
Vibe coding and spec-driven development sell the same dream: write a spec, let the AI build it, never read the code. I tried it and hit the three-month wall. Matt Pocock's talk — and his 322k-install Claude Code skills — convinced me he's right: AI makes software fundamentals (ubiquitous language, TDD, deep modules) matter more, not less. A field test against my own MCP server, AI agents, and the migration that taught me bad code is the most expensive it's ever been.
The Developer's Toolkit in 2026: Sites, AI, and Hard-Won Tips That Actually Stuck
Not a list of 100 tools. Just the developer tips, websites, and AI tools I actually use — from building production MCP servers, AI agents, and EdTech platforms in 2026.
Education
Academic background
Amity University Online
Noida, India · 2024 – 2027
Bachelor of Computer Application
GPA · CGPA 8.31 / 10
GTB Computer Education
Jalandhar, Punjab · 2016 – 2017
Diploma in Web Designing and Development
Certifications
Credentials
MongoDB Associate Developer
MongoDB
Mastering the System Design Interview
Udemy
Legacy JavaScript Algorithms & Data Structures
freeCodeCamp