Full-Stack + AI Developer · Hyderabad, India
I build things end-to-end: e-commerce with live inventory and payments, ML/AI models with measurable results, and tools that make developers faster.
| Project | What it is | Stack |
|---|---|---|
| 🌸 Arkaira | Live flower e-commerce store — realtime stock, Razorpay payments, admin dashboard | Next.js, Supabase, TypeScript |
| 🔍 GitGlance | CLI that scores any GitHub profile 0-100 with exact fixes — transparency over vibes | Python |
| 📧 SpamScope | Naive Bayes spam classifier built from scratch — no sklearn, all math readable | Python |
| 🛡️ EnvGuard | Zero-dependency typed env validation for Node — fail fast, list every error at boot | TypeScript |
| Project | What it is | Stack | Result |
|---|---|---|---|
| 🤖 AI Chatbot (RAG) | Retrieval-Augmented Generation chatbot — answers from a knowledge base with source citations + live browser UI | Flask, scikit-learn, LLM | Demo-ready |
| 🔢 Digit Recognition CNN | Convolutional Neural Network trained on MNIST, full training pipeline | PyTorch | ~99% accuracy |
| 📧 Spam Classifier | SMS spam detection with TF-IDF + Naive Bayes on real UCI dataset | scikit-learn, NLP | ~97% accuracy |
| 🎬 Movie Recommender | Content-based recommendation engine using cosine similarity | scikit-learn | Ranked results |
| 🏠 House Price Predictor | Random Forest regression with feature engineering | scikit-learn | R² = 0.80 |
| 🌸 Iris Classifier | Multi-class classification — compares 5 algorithms, decision-boundary viz | scikit-learn | ~98% accuracy |
- Building ML/AI projects end-to-end — models with measured results, wrapped in demoable products
- Deepening ML fundamentals (implementing algorithms from scratch > importing them)
- Growing my RAG chatbot with real embeddings + vector database
- Open to hackathon teams and collaborations — reach me right here on GitHub
Built with real code, honest commits, and one profile README that actually renders.