Raj
I build full-stack products., LLM pipelines., scalable backends., software that ships.
Full-stack developer with a passion for AI and systems.
About
I'm a Computer Science student at BIT Mesra building at the intersection of full-stack engineering and applied AI. I care about systems that are fast, scalable, and actually useful.
I've had the opportunity to work on real-world engineering problems, from building production-ready backend systems to working with LLM Pipelines and large-scale data. Alongside this, I built Stedzo, a multi-tenant SaaS platform that digitizes self-study libraries and now serves 500+ users across 20+ libraries.
Beyond building products, I enjoy competitive programming, with 900+ problems solved, and exploring distributed systems, databases, AI, and developer tools. I like going beyond just using technology — understanding how things work under the hood and turning that understanding into software that actually ships.
Experience
Software Engineer Intern
- Designed and shipped a semantic contract comparison pipeline — sentence embeddings, Hungarian alignment, LLM risk classification. 90% alignment recall, 77% faster analysis via async concurrency. Deployed to production.
- Architected a multi-tier subscription gating system (FastAPI, PostgreSQL, React) across 11 document tools, enabling tiered monetization.
- Shipped 6 feature enhancements and resolved critical bugs through CI/CD pipelines in Agile sprints.
Natural Language Evaluator — Hindi
- Evaluated and refined Hindi LLM outputs — diagnosed failure modes across 500+ responses, proposed structured prompt improvements that directly boosted model quality benchmarks.
- Worked cross-functionally with 5+ evaluators to identify linguistic edge cases and optimize prompt-response pipelines.
Projects
Multi-tenant SaaS for self-study libraries — Student, Library Owner, Super Admin roles. 500+ users across 20+ libraries, deployed as web app and Android app (Capacitor v6). Backend: 45 REST routes, Socket.IO chat, 11 cron jobs. 130k+ question mock test engine with KaTeX, Razorpay + OCR payment verification.
LangGraph-powered conversational agent that qualifies leads from social media conversations. Models the workflow as a state machine — nodes handle intent detection, RAG retrieval from a knowledge base, response generation, and progressive lead capture (name → email → platform). Designed for WhatsApp deployment via Meta webhooks with Redis session persistence.
Real-time IPL match prediction platform with group-based competition. JWT auth, protected REST APIs for predictions, virtual coin transactions, and live leaderboard updates with strict server-side validation. Optimized MongoDB schemas across users, groups, matches, and bets — packaged as an Android APK.
Skills
Languages
04Web & Backend
09Databases
04AI & ML
05Fundamentals
06Certifications
Generative AI Leader & Practitioner
Foundational and advanced principles of LLMs, prompt engineering, fine-tuning, and RAG architectures on Google Cloud infrastructure.
View credential ↗Blockchain Technologies & Applications
Decentralized systems, smart contract design, cryptographic consensus protocols, and Web3 applications.
View credential ↗Introduction to Philosophy
Core philosophical traditions covering epistemology, ethics, logic, and the philosophy of mind.
View credential ↗Introduction to Basic Game Development using Scratch
Game design fundamentals — event-driven programming, sprite logic, and interactive project building using Scratch.
View credential ↗Getting Started with Azure DevOps Boards
Agile project tracking with Azure DevOps — work items, sprints, backlogs, and team workflow management.
View credential ↗Achievements
Get in touch
Let's build something
worth shipping.
Open to internships, full-time roles, and interesting problems. If you're building something worth working on — or just want to talk systems and AI — reach out.