ECE student at IIT Roorkee bridging hardware and software — building intelligent agentic AI systems, applied ML solutions, and full-stack applications in fast-paced environments.
I'm Shivansh Yadav, a B.Tech student in Electronics & Communication Engineering at IIT Roorkee (CGPA 8.7). I secured AIR 3286 in JEE Advanced and AIR 3500 in JEE Mains.
My interests span full-stack web development (MERN Stack), AI/ML, and core electronics. I enjoy building real-world solutions — from placement portals with AI chatbots to hardware signal detectors.
Placement management portal built with MERN Stack. Features an AI-powered RAG chatbot for student queries on company cutoffs and automated email segregation for recruitment notifications.
Designed and fabricated a hardware receiver for the 2.4GHz band using an Envelope Detector circuit to demodulate ASK signals, with LED array visualization of the output.
Implemented a synchronous MOD-10 counter using logic ICs, interfaced with a 7-segment LED display via BCD-to-7-segment decoder. Demonstrated hardware synchronous counting.
Hybrid neural network combining CNNs (image features) and MLPs (tabular data) for real estate price prediction. Reduced MAE by 15% over baselines, processing 10,000+ listings with optimized pipelines.
AI-powered conversational honeypot (Top 2% at GUVI x HCL Hackathon). Groq LLM persona simulation with 7-layer anti-detection guardrails, multi-signal scam detection, and regex intelligence extraction for UPI IDs, bank accounts & phishing links.
DATAFORGE National Finalist (E-Summit '26): Full-stack RAG system that detects LLM hallucinations against source documents. FastAPI/LangChain backend with React UI to process atomic claims, utilizing vector search for real-time trust scores and automated correction suggestions.
Convolve 4.0 Hackathon: Local computer vision pipeline to parse highly variable loan invoices without paid cloud APIs. Extracted text, signatures, and stamps using EasyOCR, contour analysis, and regex pattern matching. Built robust fallback mechanisms for faded ink, handwritten text, and rotated documents.
Open to internships, project collaborations, and interesting engineering challenges. Feel free to reach out.
shivanshsdp@gmail.com