From AI SaaS platforms to multilingual voice assistants — production-grade systems built with modern architectures and deployed at scale.
Production applications serving real users
Full-scale AI-powered content automation SaaS with 6 FastAPI microservices, 4-tier AI fallback (GPT-4 → Gemini → Claude → Llama), Razorpay payments, and multi-platform social distribution. Serving 50+ daily users with 99.5% uptime.
Modern digital agency website with interactive animations, responsive design, and comprehensive service showcases for creative professionals.
Personal AI-powered knowledge management system built with Next.js 15. Capture notes & links, auto-summarize with Gemini AI, semantic search, and conversational Q&A against your entire knowledge base.
Intuitive note-taking app with real-time sync, rich text editor, tagging system, and clean responsive UI for enhanced productivity.
Modern product showcase and catalog website for a family electronics store. Displays products, repair services, spare parts, and store info with Supabase backend.
Intelligent systems, LLMs, agents, and computer vision
Real-time voice-first AI butler with holographic face visualizer rendered on HTML5 Canvas. English + Hindi bidirectional audio streaming with sub-second latency via Gemini 2.5 Flash Live API.
Highly capable autonomous assistant with persistent memory (pgvector), real-time bidirectional voice, camera-based perception, full computer/browser control, 110+ API integrations, and contextual habit tracking.
Simulates human-like group chat using 5 independent LLMs with distinct thinking styles — Logical, Creative, Skeptical, Practical, Synthesizer — to analyze topics from multiple angles.
Personalized mental health therapy recommendations via multi-modal sentiment analysis. Built with LangGraph agentic architecture and multi-modal input processing.
Dynamic gallery that auto-detects and classifies game screenshots using OpenAI CLIP model. Filters out non-game content intelligently with real-time WebSocket folder watching.
AI-powered data analysis agent that ingests datasets, performs automated EDA, generates visualizations, and provides actionable insights through conversational interface.
AI-powered chess engine with intelligent move evaluation, position analysis, and adjustable difficulty. Play against a machine that learns from optimal strategies.
Control your entire computer with hand and finger gestures. Real-time hand tracking with MediaPipe, mapped to keyboard, mouse, and scrolling actions.
Real-time facial recognition security system using deep learning for biometric authentication. Supports multi-face enrollment, liveness detection, and access logging.
Analyzes Twitter trends using Latent Dirichlet Allocation (LDA) topic modeling. Extracts hidden themes from tweet data and visualizes trending conversations.
End-to-end web apps with modern architectures
Professional text-to-speech and voice cloning suite powered by Gemini 2.5 Flash. Built as a Google AI Studio app with rich voice customization and real-time preview.
Full-stack YouTube to MP3/MP4 converter with React UI, Flask backend, yt-dlp + FFmpeg processing. Also ships as an Electron desktop app (Neurotube).
Prompt-to-wallpaper generator for phone and desktop resolutions. Type a description, AI generates stunning wallpapers in multiple aspect ratios.
Upload any photo and get a perfect professional headshot. AI-powered background removal, face enhancement, and professional lighting adjustment.
Web-based image background removal tool using rembg (U²-Net). Upload an image, get clean cutout. Dockerized FastAPI backend for easy deployment.
Full-stack PDF reader that reads documents aloud using natural voices. Upload PDFs up to 200 pages, select pages, customize voice/speed/pitch, and track progress.
Feature-rich web novel reading platform with chapter navigation, bookmarks, reading progress, dark mode, and clean distraction-free reading experience.
Productivity tools, automation scripts, and utilities
Real-time holographic body movement detection and visualization using computer vision. Tracks body pose, creates holographic representation overlay.
Air-type on a virtual keyboard using finger gestures. Real-time hand tracking recognizes tap gestures to input text without touching any physical device.
Control page scrolling with hand gestures using webcam. Raise or lower your hand to scroll up/down — fully touchless browser navigation.
Full-featured scientific calculator with trigonometric functions, logarithms, exponents, and expression parsing. Clean modern UI with keyboard support.
Lightweight, fast note-taking application with markdown support, local storage persistence, search functionality, and clean minimalist design.
Explore 50+ repositories on GitHub — from AI agents to full-stack SaaS platforms.
View All on GitHub