Multi‑Tenant SaaS Systems — roles, permissions, audit logs, org‑level isolation
Real‑Time Applications — WebSockets, Redis, low‑latency updates
AI‑Powered Systems — RAG pipelines, embeddings, agent workflows
Backend‑Heavy Engineering — schema design, performance, scale

Stack: Vue.js, Vuetify 2, Node.js, MariaDB

Built a production‑grade CRM for Direct Selling Agents
Fully customizable roles, hierarchy, workflows, permissions per organization
Designed and managed 65+ interlinked SQL tables for multi‑org scalability
Role‑based dashboards, detailed audit logs, leads & follow‑ups
Bulk Excel import + SSR for SEO

Closest thing I've built to a Salesforce‑level system.

Stack: React, MongoDB, Socket.IO, Cloudinary

~90% of core YouTube features implemented
Google OAuth, uploads, playlists, subscriptions, comments, likes
Real‑time notifications using WebSockets (<200ms latency)
Scalable DB design for high traffic

Stack: LangChain, LangGraph, Ollama, Pinecone / Qdrant

RAG system over Indian Central & State Acts
Custom PDF crawler, cleaner, chunker, embedding pipeline
Local SLM inference using Ollama (qwen2.5)
Contextual legal Q&A with vector search

Stack: MERN, Socket.IO, Redis

Real‑time messaging with authentication
Scalable WebSocket architecture

Stack: LangGraph, Streamlit, Gemini API

Generates full web apps from a single prompt
Live preview + downloadable source code

Engineering systems that actually scale
Clean architecture > quick hacks
AI that solves real problems, not demos

Open to internships, collaborations, and hard problems.
If it's backend‑heavy, system‑level, or AI‑driven — I'm in.