Available for full-time roles · India & remote

Raj
Shekhar

Backend Engineer & Agentic AI Builder

Final-year CS student at LPU building production AI systems — RAG pipelines, LangGraph agents, distributed microservices. Currently at NERVESPARKS engineering the future of enterprise automation.

View my work → Get in touch

3+ Production AI Systems

8.0 CGPA at LPU

8mo Engineering @ NERVESPARKS

Tech Stack

Languages

Python TypeScript JavaScript Kotlin Java C++ SQL

AI / ML

LangGraph LangChain RAG Pipelines Gemini API OpenAI API Whisper PyAnnote

Backend

FastAPI Node.js Express.js Django WebSocket REST APIs

Databases

MongoDB PostgreSQL Redis ChromaDB Weaviate MySQL

Cloud & DevOps

Docker Kubernetes GCP AWS PM2 CI/CD Terraform

Frontend

React Next.js Zustand Redux WebRTC Socket.IO

Experience

Nov 2025 – Jul 2026

NERVESPARKS · xsparks.ai
Gurugram, India

Software Engineer Intern

Architected production Agentic AI systems using LangGraph with multi-agent workflows, memory persistence, and state management — automating complex enterprise tasks end-to-end.
Built RAG pipelines with ChromaDB achieving 40% improvement in retrieval accuracy, handling 1,000+ concurrent requests at 60% reduced latency.
Designed an Audio RAG System using PyAnnote speaker diarization, OpenAI Whisper, and timestamp-based retrieval — enabling fully searchable audio knowledge bases.
Built scalable FastAPI microservices with JWT auth and WebSocket support; optimized on-device GGUF model inference in Kotlin (Iris Android App), improving response times by 35%.
Deployed containerized microservices via Docker + CI/CD on AWS ensuring zero-downtime releases.

2022 – 2026

Lovely Professional University
Phagwara, Punjab

B.Tech — Computer Science & Engineering

CGPA: 8.0 / 10 — Final year, graduating 2026.
Focus on distributed systems, system design, DSA, and microservices architecture.

Projects

01 // AI DEVOPS

AI DevOps Service Monitoring Platform

Enterprise monitoring dashboard with 100% LLM-powered anomaly detection, processing 10,000+ metrics/min. Cuts incident resolution time by 70% via automated Gemini 2.0 Flash diagnosis. Multi-tenant microservices with full data isolation for 20+ concurrent users.

React FastAPI Prometheus MongoDB Docker Gemini AI

github ↗

02 // AGENTIC AI

TerraBot — LLM-Driven Cloud Infrastructure Agent

AI-powered multi-cloud deployment agent that converts GitHub README files into production-ready Terraform configs. Reduced deployment errors by 60% and eliminates manual IaC setup. Security guardrails preventing 100% of unsafe default configs across AWS & GCP.

FastAPI LangGraph Terraform AWS GCP Gemini

github ↗

03 // FULL-STACK

Tunify — AI Music Streaming & Social Platform

Full-stack streaming platform with synchronized Party Rooms — real-time playback, WebRTC video/audio calls, live reactions via Socket.IO at sub-100ms latency for 100+ concurrent users. Gemini AI chatbot for natural language music discovery.

React WebRTC Socket.IO MongoDB Gemini Cloudinary

github ↗

04 // ANDROID AI

Iris — Offline Android AI Chat App

Offline-first Android AI chat app with local llama.cpp inference and a full RAG pipeline for document-based Q&A. On-device GGUF model inference in Kotlin — no cloud dependency, 35% faster response than baseline.

Kotlin llama.cpp RAG GGUF ChromaDB

github ↗