Projects

A collection of projects I've built and contributed to.

Developed a cross-platform mobile application focused on automated personal finance management as part of the B-Tech Final Year Project.
Ranked 1st among 70+ teams in the Mind's Eye Competition for innovation and overall product execution.
Implemented the front-end of the entire application using React Native. Features include automated transaction details extraction from SMS text, Historical statements & Analytics, Gamification elements, Budgeting, CRUD Operations for Transactions, bank accounts, and bill reminders.

React NativeNode.jsMongoDB

Built a comprehensive web application that connects patients with suitable doctors for diagnostics under Solving for India Hackathon'23 hosted by Google, AMD & GFG and secured a spot in regionals.
Worked on front-end of the application using ReactJS. Features include appointment scheduling, live video consultations, WhatsApp & email integration, disease prediction, chatbot assistance, and e-pharmacy.

React.jsFlaskMongoDB

Developed a web application that simplifies and streamlines summarizing team meetings in just 4-5 minutes under Mastek's Project Deep Blue Hackathon, securing a spot in the Semifinals.
Application supports various file formats such as text, audio, and video. Developed and integrated a Chrome extension with Google Meet to provide live captioning, cloud storage, and highlighter.

React.jsDjangoMongoDB

Implemented front-end of a comprehensive platform for coding & interviews with a focus on peer learning.
Developed various features including peer learning using cohort models based on ratings derived using feedback loops for interview preparation.
Other key features include a code runner using JDoodle API, a live interview with Jitsi, a collaborative code editor using SocketIO, learning modules, and user dashboard & analytics.

Next.jsNode.jsMongoDB

Built a local reverse proxy middleware for LLM APIs that enables real-time prompt sanitization and content filtering, detecting and redacting sensitive data (PII, API secrets, credentials) before external transmission.
Implemented intelligent safeguards specifically designed for AI agent workflows, preventing accidental leakage of API keys, authentication tokens, database credentials, and confidential business data in developer environments.
Developed pattern-based detection system for identifying and masking various sensitive information including SSNs, credit card numbers, API keys, tokens, and custom regex patterns with configurable redaction strategies.

PythonTypeScript

Trace code changes back to the prompt that produced them — GitLens, but for prompts.
PromptTrace captures every change a coding-agent prompt makes to your files and ties it back to that prompt, so you can answer which prompt touched a file and exactly what it changed.
The extension adds a prompt diff view, a timeline of prompts that modified the active file, and blame-style attribution with gutter markers and hover details for line-level context.
It also supports reindexing after exclude-rule changes, retention controls for storage management, and automatic capture for Claude Code via hooks.

TypeScript