S.YOGADHARSHAN

Full-Stack Developer | AI Engineer | React Developer | Prompt Engineer

Portfolio: yogadharshan.vercel.app

Phone: +91 63791 35594 | Email: yogadharshan3@gmail.com | Rajapalayam, Tamil NaduLinkedIn: linkedin.com/in/yoga-dharshan-30192a30b | GitHub: github.com/MRdivine17

Summary

Full-stack developer and AI engineer shipping production software since 2025; third-year B.Tech AI and Data Science student. Builds React and Next.js apps, Node.js and Convex backends, and LLM features with prompt chains, RAG and LangGraph. Sole developer of RIT Canvas, an LMS used by 2,139 students and 169 educators.

Education

Ramco Institute of Technology

Rajapalayam, Tamil Nadu

B.Tech, Artificial Intelligence and Data Science; CGPA: 7.13/10

2024 - 2028

Experience

Student Software Developer

2025 - Present

Ramco Institute of Technology

Rajapalayam

  • Architected and solely built RIT Canvas (ritcanvas.in), a Next.js and React Native LMS serving 2,139 students and 169 educators within 5 months of the first commit: 55 tables, 302 functions.
  • Migrated RIT Canvas off Vercel to a 28-core server, scaling from 1 to 10 Docker replicas with Dokploy, Traefik, Cloudflare Tunnel and Convex.
  • Engineered LLM pipelines from teachers' PDF, DOCX and PPTX files: quizzes from 12,000-character chunks with prompts tuned for type rotation and top-ups, and study notes from a six-pass prompt chain.
  • Delivered exams, live invigilation, grading and analytics for the two-developer RIT REACH platform (ritreach.ritrjpm.ac.in): 8 tables, about 10,000 lines.
  • Co-developed the 14-module RIT ERP in Django as one of 7 developers.

Lead Developer (Freelance)

2025 - Present

Solar Scripts (solarscript.in)

  • Shipped 10+ premium FiveM (GTA V) scripts and 50+ custom commissions for a 500+ member community, in Lua, React and MySQL that auto-detect ESX, QBCore and Qbox at runtime.
  • Cut Solar CarPlay's React UI CPU time by 80% (532 to 105 ms/s in DevTools) by swapping full-screen CSS filters for overlays and throttling animations.
  • Designed a 35-table MySQL schema and 151 callbacks for a 17-page police dispatch system.

Organiser and Developer

Sep 2026

HackFinity 3.0, Ramco Institute of Technology

  • Co-organised a 24-hour hackathon for 114 participants and built its React Native judging app (35 screens, 5 roles), cutting the APK from 108 to 61 MB.

AI Intern

Jun 2026

Mentron

  • Implemented embedding search and structured LLM output with LangChain.

MERN Stack Intern

Dec 2025

Notasco Technologies India Pvt Ltd

Madurai

  • Developed a MERN e-commerce app in 1 month with JWT auth and REST APIs.

Projects

RIT Pulse | Next.js, Convex, TypeScript

pulse.ramcoad.com

  • Launched a live quiz platform for up to 1,000 players, designed to save 90,000 database reads per question at 300 players by reading a stored count.

LUMEN query agent | Python, LangGraph, Gemini, FastAPI, PostgreSQL, MongoDB

  • Orchestrated a five-node LangGraph agent turning plain-English questions into read-only SQL or MongoDB queries, with ChromaDB RAG schema lookup and regex plus LLM guards.

Technical Skills

  • Languages: TypeScript, JavaScript, Python, SQL, Lua
  • Frontend and Mobile: React, Next.js, React Native, Expo, Tailwind CSS
  • Backend: Node.js, Express.js, Convex, Django, FastAPI, REST APIs, JWT, RBAC
  • AI and LLM: Prompt engineering, LangChain, LangGraph, RAG, embeddings, Gemini, Ollama, ChromaDB
  • Databases: PostgreSQL, MySQL, MariaDB, MongoDB, Supabase, Firebase
  • DevOps: Docker, Dokploy, Traefik, Cloudflare Tunnel, Vercel, CI/CD, Git