AI-native web apps
End to end — Angular or React front, Node or NestJS back, LLM features built in rather than bolted on after. Built to run in production, not to impress in a slide deck.
Available for select freelance projects
Dhruv
Parekh
Full-Stack & AI Engineer
Mumbai, India · available worldwide
You ship it. I build it. Seven years of production web apps on the MEAN stack — Angular is home. Now I build the AI layer too: RAG, agents, LLMs wired into apps that actually ship. Most devs do one or the other. I take on select freelance projects that need both.
Who is Dhruv Parekh?
Seven years shipping real products across five companies. MEAN stack by trade — Angular is home, TypeScript everywhere, Node and Express under the hood. Currently Senior Frontend Developer at Volofin, building fintech in Mumbai. The kind of code that handles other people's money — so it has to be right. The day job is live, not a gap to fill.
The rare part: I went deep on AI instead of stopping at an API call. LangChain, LangGraph, RAG pipelines, vector databases, fine-tuning — Python alongside TypeScript. Most web devs can't build this. Most AI people can't ship the app around it. I do both, which is the whole point. The gap between web dev and AI is smaller than people think.
On the side I take on select freelance projects — the ones that need a senior engineer who can architect the system, write the frontend, and wire the AI in without it falling over in production. Day job stays; the good projects still get in. The hard ones are more interesting.
MEAN
Core Stack
AI
Current Focus
Fintech
Domain
MUM
Based In
What I build, freelance
Select projects on the side — the ones that need both halves: a production web app and a real AI layer inside it.
RAG & retrieval systems
Chat over your own data. Document pipelines, vector search on Pinecone or ChromaDB, grounded answers with real citations — not a chatbot that guesses.
LLM agents & automation
Multi-step agents with LangGraph and the OpenAI or Anthropic APIs — tools, memory, and guardrails that hold up under real traffic.
Fix the backend that's buckling
Slow APIs, tangled code, an architecture that can't scale. System design, SOLID, event-driven — the patterns that keep a codebase shippable.
What I work with
The tools I reach for on client work. MEAN stack is home turf; AI is where I'm pushing now.
Frontend
- Angular
- TypeScript
- JavaScript ES6+
- RxJS
- NgRx
- HTML5
- CSS3 / SCSS
- Bootstrap
- Angular Material
- React
Backend
- Node.js
- Express.js
- NestJS
- REST APIs
Database
- MongoDB
- Mongoose
- PostgreSQL
- MySQL
- Redis
- Firebase / Firestore
- Supabase
DevOps & Cloud
- Git / GitHub
- Docker
- AWS
- CI/CD
- GitHub Actions
- Nginx
- Linux / Bash
Architecture & Practices
- System Design
- SOLID Principles
- Design Patterns
- Agile / Scrum
- Event-Driven
- Cypress
AI & Machine Learning New
- LangChain
- LangGraph
- OpenAI API
- Anthropic / Claude
- RAG
- Vector Databases
- Hugging Face
- Ollama
- Prompt Engineering
- Python
- LLM Fine-tuning
- Pinecone / ChromaDB
Got a project that needs both halves?
A production web app and a real AI layer inside it. That's the work I take on. Tell me what you're building — an idea, a half-finished app, or an AI feature you can't get over the line. I'll tell you straight whether I can ship it.