

LiveWeb app · AI
whatsforyou
Upload a job offer letter and get a straight answer. It flags nine kinds of hidden traps, quotes the exact clause and works out your real monthly in-hand pay.
- Next.js
- TypeScript
- Gemini
- Vercel
Apps people can open in a browser today, from AI tools to full-stack platforms.
A research prototype for cognitive assessment: browser-based tasks, consent-based speech transcription and saved reports. Not a diagnostic tool.
React · Express · PostgreSQL
Ten GRE questions that adapt to each answer using Item Response Theory, then an AI-written study plan.
FastAPI · MongoDB · Gemini
Browse, search and buy stock, with admin tools for the lot. Built test-first on a JWT-secured API.
React · Express · MongoDB · Vitest
Checks a loan offer against registered lenders, rates and app permissions, to help people avoid predatory lenders.
JavaScript · Python
Order management for a dry cleaner: automatic billing, status tracking and a live business dashboard.
React · Express · SQLite
An online store where shoppers place 3D furniture in their own room using AR on their phone.
WebAR · 3D models · JavaScript
A high-concurrency consultation backend with a booking state machine, distributed locks and a full CI pipeline.
Express · PostgreSQL · Prisma · Redis
Flutter and React Native apps, including one live on Google Play.
Legal awareness for students: scenario quizzes, step-by-step FIR guidance and a lawyer connect feature.
Flutter · Firebase
Mock entrance tests for the Army Institute of Law, with 200 Gemini-generated questions and timed sections.
Flutter · Gemini
Browse a sports turf, pick a date and time, and manage your bookings.
React Native · Expo
An offline-first task manager with priorities, reminders and sync when back online.
Flutter · BLoC · SQLite
A two-role request and confirmation flow with item-level approval and partial fulfilment.
Flutter · Riverpod
Agents that listen, search, reason and act, built with real tool calling and retrieval.

Turns pharmaceutical complaints from PDFs into structured records, assesses risk and lets QA staff correct fields by chatting.
React · FastAPI · LangGraph · PostgreSQL
A provider-agnostic chat platform with streaming, session memory, rate limits and moderation.
MongoDB · Express · React · Node
Agents that gather the day’s US market news, summarise it, translate it into four languages and post to Telegram.
CrewAI · Groq · Telegram
Automates dataset onboarding from Google Drive: metadata, suggested quality rules and a catalogue entry.
FastAPI · MCP · Drive API
Pulls structured data out of PDFs with GPT-4o, shows it in tables and exports to Excel, CSV and JSON.
Flask · OpenAI · Textract
Schedule meetings in plain English. It finds free slots and sends Google Meet invites.
Python · Gemini · Google Calendar
A Chrome extension with a rule-based planner, semantic and keyword search, and a full reasoning trace.
JavaScript · Chrome extension
Models trained, evaluated and put to work on real problems, including two published papers.

Move files between devices with hand gestures, using real-time hand tracking and sockets. Published as a research paper.
MediaPipe · Python · Sockets
A CNN that verifies genuine ISI certification marks on helmets. Published in the IJCRT journal.
OpenCV · TensorFlow · Keras
A deep learning risk model trained on three public datasets, with a Flask web interface.
TensorFlow · Flask
A reproducible benchmark of Whisper on LibriSpeech, reporting WER, CER and latency.
Whisper · Hugging Face
An interactive dashboard showing what drives placement outcomes, with filters and SQL aggregation.
Streamlit · Plotly · SQLite
Bots and pipelines that take repetitive work off people’s plates.
Fills PDF forms by finding labels and blank regions with OCR and computer vision. No coordinates to configure.
Tesseract · OpenCV · Python
Checks uploaded transaction files for bad phone numbers, dates, duplicates and formats, then splits the output.
FastAPI · Streamlit
Finding what’s wrong before it reaches your users.
Project notes
Problem: People reviewing job offer letters need to identify clauses and understand the pay described in the offer.
What was built: A web app that flags clauses, quotes the relevant text, and estimates monthly in-hand pay.
Technology: Next.js, TypeScript, Gemini, and Vercel.
Product behavior: The app analyzes an uploaded offer letter and presents flagged clauses and an estimated monthly pay figure.
Problem: Students need to order from campus cafes and keep track of their order token.
What was built: A Flutter app for cafe discovery, pre-ordering, UPI payment, and token tracking, alongside order management for shop owners.
Technology: Flutter, Firebase, Firestore, and UPI payments.
Product behavior: Students can place orders and follow token progress; shop owners can manage orders in real time, as described in the project listing.
Problem: Finding a relevant Drive file can be difficult when a person does not know the exact file name or folder.
What was built: A RAG and tool-calling assistant that turns plain-English requests into Google Drive queries.
Technology: LangChain, FastAPI, Streamlit, and Google Drive API.
Product behavior: A person can ask for files in natural language and the assistant searches Drive for relevant items.
Tell us what you have in mind. We reply within 24 hours.