Full Stack Engineer. I wire up LLMs to real workflows, build Salesforce integrations, and ship products people actually use — not just demo.
A mix of SaaS, automation, and tooling. Every project shipped with real users in mind.
Writing test cases manually from a codebase is slow, inconsistent, and the first thing skipped under deadline pressure.
Full-stack platform that connects to a GitHub repo via OAuth 2.0, parses the AST, and uses Google Gemini 2.3 Pro to auto-generate full test suites — posting results as automated PR comments.
Generic resume templates produce generic resumes — job seekers struggle to articulate their experience in a way that stands out.
Real-time collaborative resume editor powered by OpenAI GPT-4. Adaptive prompt templates generate tailored content per role, with live PDF export via Puppeteer.
URL shorteners are either too simple (no analytics) or too bloated (enterprise pricing for basic needs).
Secure URL shortener with Base-62 encoding, JWT + refresh token rotation, real-time Chart.js analytics, and sub-100ms response latency handling 1,000+ monthly redirects.
Repetitive LinkedIn outreach is time-consuming and most automation tools are easily fingerprinted and blocked.
Go-based browser automation using Chrome DevTools Protocol with anti-detection: fingerprint randomization, Bézier-curve mouse simulation, exponential backoff, and a modular JSON-configurable plugin architecture.
Teams managing tasks across tools lose context — no single view of ownership, status, and search in one place.
Full-stack task management app with JWT authentication, CRUD operations, real-time search and status filtering, and a responsive React dashboard backed by a Node.js/MongoDB API.
To-do lists grow endlessly — users have no objective way to decide what to work on next.
Django-based scoring engine that ranks tasks by urgency, importance, effort, and dependencies. Four sorting strategies, circular dependency detection, and a clean JS frontend with API endpoints returning human-readable explanations per task.
Manually placing Market, Limit, and Stop-Limit orders on crypto exchanges is error-prone and slow during volatile conditions.
Python trading bot with HMAC-SHA256 signed REST API calls to Binance Spot Testnet. Supports all three order types via both a CLI (argparse) and a cyberpunk-themed Flask web UI with structured file + console logging.
Free online tools are scattered across dozens of sites — each requiring a sign-up, slow to load, and sending your data to a server.
Client-side utility toolkit for images, math, finance, and SEO — everything runs in the browser with zero data transmission. Built with Next.js 14 and Tailwind, deployed on Vercel.
Engineering teams have no visibility into how much they're spending on SaaS and AI tooling — costs sprawl silently.
AI spend optimization platform that audits SaaS and AI tooling costs, identifies redundant subscriptions and savings opportunities, and generates actionable recommendations. Built with Next.js and TypeScript.
Manually checking compliance documents against rules is tedious, inconsistent, and doesn't scale.
Full-stack app that uploads PDFs and evaluates them against 3 user-defined natural language rules using GPT-4o-mini. Returns pass/fail status, evidence, reasoning, and confidence score per rule — with a premium dark UI and real-time feedback.
Growing an Instagram account organically requires hours of daily manual engagement across hashtags — most people simply don't have the time.
React Native (Expo) mobile app where users connect their Instagram account, pick interest categories, and a Python/Flask automation server runs scheduled background sessions — browsing hashtags and engaging with content every 30 minutes.
Golf clubs running charity fundraisers have no purpose-built platform — they manage subscriptions, scores, and draws manually via spreadsheets.
Full-stack subscription SaaS where golfers pay monthly via Stripe, submit Stableford scores through a verification flow, and enter automated monthly prize draws. A share of every subscription is routed to charity campaigns. Includes an admin panel and transactional emails via Resend.
Students juggle notes, schedules, and resources across disconnected tools with no intelligent layer tying them together.
AI-powered study operating system built with Next.js and Supabase. Features a modular architecture with dedicated services, stores, hooks, and middleware — designed to centralize study workflows with AI assistance.
Photography businesses manage jobs, edits, and client deliveries across spreadsheets and email threads with no structured workflow.
Production-grade photo management system with a FastAPI backend, PostgreSQL via Alembic migrations, background workers for async processing, Docker deployment, and a web frontend — built around a specification-driven architecture.
Reusable UPI payment QR codes can be scanned by anyone — there's no way to issue a one-time payment link that expires after use.
Node.js/Express app that generates token-based UPI QR codes. Each token is deleted server-side after the payer confirms payment, making the link permanently unusable. Supports direct UPI QR and a permanent merchant QR mode. Deployed on Render.
There's no clean, searchable interface for Bangalore's 100+ pincodes — lookups require knowing the exact pincode or trawling government sites.
Full-stack React + Node.js app with bidirectional search (pincode ↔ area name), zone filtering (Central/North/South/East/West), and a modal detail view. REST API with full error handling and input validation.
Training institutes track student placements in spreadsheets — no portal for job listings, student profiles, or placement reporting.
Training and placement management website built with Python/Django and a rich SCSS frontend. Handles student profiles, job listings, application tracking, and placement reporting in a unified platform.
A full stack engineer wiring AI to real workflows. Building products that ship and survive production.
// AI-Native Full Stack Developer — 2 internships, 15+ projects shipped
// LLM integrations, SaaS, Salesforce Apex, browser automation
// I use AI agents to build faster — I own every line that ships
MCA graduate with hands-on internship experience at A2IT Online and Draftsstech. I build production-grade software across the stack — React/Next.js frontends, Node.js/Django backends, LLM integrations via OpenAI & Gemini, and Salesforce platform development with Apex and LWC. I use AI coding agents as a core part of my workflow — not just for autocomplete, but for architecture, testing, and shipping. 15+ projects shipped to production.
I dig past the surface ask. The right architecture comes from understanding the user's actual workflow.
Working software in days, not months. Iterate on real signal — not on imagined edge cases.
Fast UIs and lean APIs are not a polish step — they're the baseline for a product worth using.
I pick boring, proven tech for foundations and modern tools where they earn their place.
2 internships, freelance, and 15+ independently shipped products — as an AI-Native Full Stack Developer.
Draftsstech — Remote
Stipend-based remote internship. Built responsive React.js component libraries with Tailwind CSS, designed RESTful APIs with Node.js/Express, implemented Redis caching and lazy loading cutting page load times by 40%, and shipped features in 2-week Scrum sprints with full Jest test coverage.
Freelance & Independent
Shipped 15+ production projects using AI coding agents as development partners — AI TestCase Generator (Gemini + GitHub OAuth), AI Resume Builder (GPT-4 + Socket.io), Golf Charity Platform (Stripe + Supabase), FeedFlow (React Native + Flask), LinkVault (1,000+ monthly redirects, sub-100ms), and Go-based LinkedIn automation with Chrome DevTools Protocol.
A2IT Online — Chandigarh
Built multi-tiered Django/DRF applications, integrated JWT + OAuth 2.0 auth, deployed to AWS EC2 via GitHub Actions CI/CD, and refactored legacy jQuery to React — reducing bundle size by 40%.
Baba Ghulam Shah Badshah University · CGPA 8.5/10
Specialization in Full Stack Development, Cloud Computing, and AI. Capstone: AI-Powered Resume Builder using React.js, Node.js, and OpenAI API.
Government Degree College, Ganderbal
Foundation in DSA, DBMS, OS, and Computer Networks. Final project: Online Attendance Management System in PHP/MySQL.
“Clean code, solid architecture, zero hand-holding needed. Delivered production-ready work consistently.”
“Rare to find someone who understands both the LLM layer and the infrastructure. He owns the full stack.”
“The automation system he built handles edge cases I hadn't even thought of. Thoughtful engineer.”
“His Salesforce Apex work saved us weeks of manual effort. Flows, triggers, LWC — he handled it all.”
“The resume builder he built got real user traction fast. Solid product instincts, not just an engineer.”
“Asked for a URL shortener. Got sub-100ms responses, analytics dashboard, and JWT auth out of the box.”
“He diagnosed our Redis bottleneck in one session and cut our p95 latency by 60%. Highly recommend.”
“Clear communicator, ships fast, and actually reads the brief. Refreshing to work with.”
“Clean code, solid architecture, zero hand-holding needed. Delivered production-ready work consistently.”
“Rare to find someone who understands both the LLM layer and the infrastructure. He owns the full stack.”
“The automation system he built handles edge cases I hadn't even thought of. Thoughtful engineer.”
“His Salesforce Apex work saved us weeks of manual effort. Flows, triggers, LWC — he handled it all.”
“The resume builder he built got real user traction fast. Solid product instincts, not just an engineer.”
“Asked for a URL shortener. Got sub-100ms responses, analytics dashboard, and JWT auth out of the box.”
“He diagnosed our Redis bottleneck in one session and cut our p95 latency by 60%. Highly recommend.”
“Clear communicator, ships fast, and actually reads the brief. Refreshing to work with.”
Whether it's a SaaS MVP, an AI integration, or a full-time role — I'd love to hear about it.
