I'm Rishab, a full-stack developer who designs and ships end-to-end systems — from internal tools and business platforms to applied AI — built for production, not just to impress in a demo.
I'm a full-stack developer based in Bengaluru, currently completing my Master's in Computer Applications. Rather than working through tutorials, I've spent the last few years building things people actually use — a full ERP platform for a manufacturing company, an AI-powered tool that replaces a mouse with hand gestures and voice. My toolkit centers on JavaScript, Python, Node.js, Express.js and MongoDB, but the real focus is always the same: software that works cleanly end-to-end, from what a user sees down to how the data is stored. If you have something that needs building and shipping, that's where I come in.
I work across the entire stack, so a project keeps moving instead of stalling while pieces wait on each other.
Two different problems, two full solutions — a business system a company depends on daily, and a hands-free AI tool built from the ground up.
A full-stack ERP system built for a manufacturing company, covering invoicing, purchase orders, customer and product management, production tracking, payment tracking, route cards, employee management, and analytics — built as a college project for a real organization's workflow.
A full-stack web application built to digitize and centralize the day-to-day business and manufacturing operations of R.P. Industrial Components. Customers, products, purchase orders, invoices, payments, production, route cards, employees, and analytics live in one platform. It was designed around the company's actual workflow, with interconnected modules rather than isolated CRUD pages, so information entered once is reused everywhere it's needed.
Replace fragmented, manual processes with one digital system that manages operational data efficiently and gives better visibility into finances and production, across three areas:
This lets the ERP represent real business processes instead of a collection of independent pages.
Built from live operational data (not static values): total revenue and expenditure, payments received and made, pending incoming and outgoing payments, top product, top customer, and company growth — with charts for revenue vs. expenditure, production distribution, and pending payments.
Client–server architecture with a modular backend split into dedicated routes for invoices, customers, purchase orders, payments, products, production, route cards, employees, and authentication.
Running as a live, cloud-hosted system rather than only a local project.
I worked across the complete development lifecycle:
R.P. Industrial Components now has a centralized platform for its major business and manufacturing workflows, reducing repetitive manual work, organizing operational data, and giving clearer visibility into finances and production. The project reflects hands-on work in full-stack development, REST API design, database integration, process automation, document generation, data visualization, cloud deployment, and PWA development.
A college project enabling touch-free computer interaction using hand gestures and voice commands, combining computer vision with speech recognition for hands-free control.
A touch-free human–computer interaction system that lets you control a computer with hand gestures and voice commands instead of a physical mouse and keyboard. A webcam captures hand movement in real time, MediaPipe tracks hand landmarks, and Python with OpenCV turns recognized gestures into actions like cursor movement, clicking, scrolling, and dragging. A separate voice module handles commands such as slide navigation and clicks. Everything runs locally on the computer's own webcam and microphone, making it a low-cost, contactless alternative.
Runs alongside the gesture system and uses speech recognition with PyAutoGUI to execute commands at the operating-system level, including:
Voice-driven cursor movement was intentionally left out because continuous voice movement added delay and hurt responsiveness.
Commands are identified from combinations of finger states, using the detected finger positions rather than any hardware buttons.
| Gesture | Action |
|---|---|
| Index finger | Cursor movement |
| Closed fist | Click / drag interaction |
| Index + pinky | Right click |
| Index + middle | Scroll up |
| Index + middle + ring | Scroll down |
| Four fingers | Application switching |
| Middle + ring + pinky | Lock / unlock |
Built as modular components: hand detection, landmark tracking, gesture recognition, cursor control, mouse actions, voice control, system control, and a feedback / status module.
MediaPipe returns landmark coordinates for the hand, and the application checks whether each finger is extended or folded. Those finger states map to predefined gestures. For the cursor, the index finger's webcam position is converted to screen coordinates and smoothed for natural control. Once a gesture is recognized, PyAutoGUI and Autopy carry out the action, while the voice module runs independently alongside.
The complete interaction pipeline: webcam hand detection, landmark tracking, finger-state and gesture recognition, gesture-to-action mapping, smooth cursor movement, clicks, drag-and-drop, scrolling, zoom and navigation, Alt + Tab switching, cursor lock, voice command recognition, presentation control by voice, on-screen feedback, and the separation of gesture and voice modules.
Shows how computer vision and voice recognition can combine into a practical touch-free interface, turning natural hand movements and simple voice commands into real-time computer actions. It also built hands-on experience in real-time computer vision, gesture recognition, human–computer interaction, automation, multithreading, and combining multiple input types in one application.
What I actually reach for day to day, grouped by where it's used.
What the user sees and interacts with.
Where the logic and data actually live.
How I build, ship, and keep things stable.
The part of the job that isn't code.
Where I first put these skills to work inside a real team.
Built responsive web applications end to end using HTML, CSS, JavaScript, and Firebase — shipping frontend features, wiring up Firebase for backend and data handling, and debugging issues under real project deadlines. It's where I learned to work within someone else's codebase and someone else's timeline, not just my own.
The formal side of how I got here.
Building on my undergraduate foundation with deeper, more advanced software development coursework.
Finished with a 9.08 CGPA and the Best Outgoing Student award — but the more useful part was everything built alongside the coursework, including both projects on this page.
Coursework and training I've completed alongside my degree.
Common questions, answered instantly — click to expand.
Reach out on any of these.