Crafting fluid algorithmic interfaces and deep machine learning architectures. Moving beyond generic templates to deliver highly sophisticated, humanized web interactions that blur the boundary between software and art.
Engineering
Digital Fluidity.
I am an architect of web systems deeply rooted in React, Next.js, and scaling robust APIs. My approach strips away rigid templates, focusing instead on delivering custom, high-fidelity experiences that adapt natively to their constraints.
By fusing raw technical competency with an editorial design philosophy, I engineer systems that don’t just process data—they narrate it.
Experience
Hackathon Participant
Active participant in various hackathons, building innovative solutions under time constraints.
Independent Developer
Building and shipping self-driven projects, focusing on AI-powered platforms and automation tools.
Founder, Vitamend
Leading the development of an AI-driven platform to streamline medicine donation and redistribution, focusing on accessibility, safety, and scalable healthcare impact.
Database Management System Intern For Swiggy
Contributed to the management and optimization of enterprise-grade databases powering Microsoft applications and Salesforce CRM workflows.
Selected
Works.
Vitamend
Medicine waste and lack of donation infrastructure
A medicine donation platform with AI-based verification and expiry detection, connecting donors with those in need while ensuring medication safety.
- AI-powered medicine verification
- Expiry date detection via image recognition
- Donor-recipient matching algorithm
- Real-time inventory tracking
Vocalis
Need for intelligent voice-driven AI assistants in web applications
An AI-powered virtual assistant and experimental product platform that leverages natural language processing for seamless human-computer interaction.
- Voice recognition and synthesis
- Context-aware conversations
- Plugin architecture for extensibility
- Real-time response streaming
DocEasy
Scattered document storage and inefficient file size management
DocEasy is a centralized document management platform that allows users to securely store all important documents and IDs in one place. The platform enables smart document compression and flexible downloads, allowing users to retrieve files in the required resolution or format without maintaining multiple copies.
- Centralized storage for documents and personal IDs
- Intelligent document and image compression
- Download files in custom resolution or format
- Secure access and organized document categorization
- Cloud-based storage with fast retrieval
3D Model Generator
Creating 3D models requires specialized skills, tools, and significant time, making it inaccessible for non-designers and rapid prototyping
3D Model Generator is an AI-powered platform that allows users to generate 3D models from simple text prompts or inputs. It simplifies the 3D creation process by eliminating the need for complex design software, enabling developers, designers, and creators to quickly prototype and visualize ideas in a 3D environment.
- Generate 3D models using simple text prompts
- Real-time 3D preview and interaction
- Export models in standard formats (GLTF/OBJ)
- User-friendly interface for non-designers
- Fast prototyping for developers and creators
- Cloud-based storage and model management
Kolkata Metro Route Planner
Planning metro journeys efficiently is difficult due to multiple lines, interchanges, and lack of clear travel time estimation in a single interface
Kolkata Metro Route Planner is a modern, interactive web application that helps users find the most efficient routes across the Kolkata Metro network. It uses Dijkstra's algorithm to compute optimal paths while providing detailed travel insights like time breakdowns, interchange penalties, and peak-hour adjustments, all within a visually intuitive interface.
- Smart route calculation using Dijkstra’s shortest path algorithm
- Interactive metro map with clickable stations
- Support for all major Kolkata Metro lines
- Detailed travel time breakdown including dwell and interchange time
- Peak hour mode with adjusted time estimates
- Option to minimize interchanges for convenience
- Dark and light mode support
- Fully responsive and mobile-optimized interface
WelcomeToMyCity
Urban travel and tourism planning is fragmented across multiple apps, making it difficult to compare transport options, explore attractions, and plan efficient itineraries in one place
WelcomeToMyCity is a comprehensive civic-tech platform that unifies city transport planning, tourism discovery, and itinerary generation into a single seamless experience. It enables users to compare multiple transport modes, explore curated city attractions, and generate optimized multi-day travel plans across major Indian metro cities.
- Multi-modal transport planner (Metro, Bus, Taxi, Walking)
- Real-time cost and travel duration estimation
- Comprehensive bus route database with smart route matching
- City explorer with categorized tourist attractions
- Search and filter functionality for places
- AI-based itinerary generation for 1–3 day trips
- Route optimization for efficient travel planning
- Multi-city support across 12 major Indian metro cities
- Google Maps integration for navigation
- Dark and light mode support
Restaurant Reservation Chatbot
Booking restaurant reservations is often time-consuming and fragmented, requiring users to navigate multiple platforms instead of using a simple conversational interface
Restaurant Reservation Chatbot is a conversational AI-driven web application that allows users to seamlessly book restaurant tables through natural language interaction. It simplifies the reservation process by providing real-time availability, booking confirmations, and menu-related queries within an intuitive chat-based interface.
- Conversational AI interface for booking reservations
- Real-time table availability checking
- Instant booking confirmation with reference ID
- Menu information and query handling
- Responsive design for mobile and desktop
- Session reset and conversation control
- Scalable architecture with modular documentation
khabri.ai
The rapid spread of misinformation and fake news makes it difficult for users to verify the authenticity of headlines and social media forwards in real-time
khabri.ai is an AI-driven news verification platform that combats misinformation by cross-referencing claims against trusted global and local sources. It provides instant, evidence-backed verdicts using real-time search data, enabling users to verify news credibility with transparency and confidence.
- AI-powered fact-checking for headlines and social media content
- Real-time news feed from trusted global and Indian sources
- Evidence-backed verdicts with source links
- Verdict classification (TRUE, FALSE, MISLEADING, NOT VERIFIED)
- Multilingual support (English, Hindi, Bengali)
- Premium UI with smooth animations and responsive design
- Dark and light mode support
PRIVSIGHT
Sensitive on-screen information is vulnerable to shoulder surfing and unauthorized viewing, with limited real-time privacy protection solutions available on desktop systems
PRIVSIGHT v2.0 is a cross-platform desktop application that uses real-time facial recognition to intelligently protect user privacy. By detecting the presence of authorized and unauthorized viewers through the webcam, it dynamically controls notification visibility, ensuring sensitive information remains secure without interrupting the user experience.
- Real-time facial recognition using webcam input
- Automatic notification blurring for unauthorized or multiple viewers
- Local processing with no cloud dependency for enhanced privacy
- Manual privacy mode toggle for instant screen protection
- Configurable sensitivity, detection frequency, and blur intensity
- System tray integration for always-on monitoring
- Cross-platform support for macOS and Windows
- Modular architecture with Electron frontend and Python backend
Additional Engagements
AIR MOUSE
Webcam Based Air Mouse Controller
Python-based air mouse controller using hand gestures for real-time cursor movement and control, enabling touchless and intuitive human–computer interaction.
- Hand Tracking - Real-time 21-point hand landmark detection using MediaPipe
- Gesture Recognition - Detect pinch, fist, pointing, open palm, peace sign, thumbs up
- Cursor Control - Move mouse, click, drag, and scroll with hand gestures
- Air Drawing - Draw in the air with your finger and save your artwork
- Voice Commands - Optional voice control for hands-free actions
- Kalman Filtering - Smooth, jitter-free cursor movement
Gesture Media Controller
Webcam based media controller using hand gestures
Python-based AI gesture controller enabling intuitive hand-gesture interactions. Supports real-time gesture recognition, responsive controls, and easy integration with devices and applications for seamless user experience.
- Real-time hand tracking using Google MediaPipe
- Intuitive gesture recognition for media control
- Visual feedback with heads-up display overlay
- FPS monitoring for performance tracking
- Gesture cooldowns to prevent accidental repeated triggers
- Smooth motion detection using moving average filtering
Arsenal & Utilities
TypeScript
React
Next.js
Python
Node.js
PostgreSQL
Tailwind CSS
AWS
Docker
GitTypeScriptReactNext.jsPythonNode.jsPostgreSQLTailwind CSSAWSDockerGitContact
If you'd like to discuss a project or just say hi, I'm always down to chat. Feel free to reach out through any of the channels below.