AI / ML Developer·Full-Stack Builder·Vibe Coder
I build intelligent systems that actually work — from deep learning pipelines to web apps that feel alive. If it involves AI, I'm probably already obsessed with it.
About Me
Hey! I'm Uthkarsh — a AI and DS student at Amrita Vishwa Vidyapeetham who genuinely gets excited about AI. Like, embarrassingly excited. The kind where I'll lose an entire weekend just to understand why a model is making weird predictions.
I got into ML because I wanted to build things that could think and adapt, not just execute instructions. These days I'm busy training models, building full-stack apps, and exploring whatever cool new thing just dropped in the AI domain.
When I'm not coding, I'm reading a research paper, messing with UI design, or vibe coding something random at 2am that somehow turns into a real project. That's basically the origin story for half my GitHub repos.
What I Work With
What I've Built
Intelligent web assistant that generates personalized study plans, creates quizzes using ML difficulty classification, summarizes long texts with extractive NLP, and delivers AI-powered study tips. Built on Flask + Python with TensorFlow, NLTK & Gensim powering the backend.
Deep learning-powered data aggregation engine using neural attention mechanisms to consolidate multi-source data streams and extract high-level insights — designed for real-world data fusion tasks.
Retrieval-Augmented Generation chatbot grounding LLM responses in real documents. Feed it any knowledge base, get accurate answers with no hallucinations — just facts from your data.
Smart scheduler resolving complex timetabling constraints automatically. Input subjects, faculty, and rooms — the algorithm handles conflicts and outputs an optimal schedule.
Full-featured platform for creating, managing, and tracking events end-to-end — handles registrations, attendee lists, scheduling, and real-time updates, built for scale.
Collection of ML experiments — classification, regression, clustering — organized across branches to compare model architectures, datasets, and training strategies systematically.
Autonomous DIY robot built with Arduino and IR sensors. Implements a PID control loop to navigate complex black tracks with high precision and speed.
Structural analysis of a Warren truss bridge using ANSYS. Conducted FEA to visualize stress concentration, deformation, and factor of safety under various loading conditions.
Let's Connect
Whether it's a collab, an internship, a project idea, or you just want to nerd out about AI — my inbox is always open. Let's build something cool together.