I'm an MCA student at AIMIT, St Aloysius (Deemed to be University), Mangaluru. I work across full-stack JavaScript, Python back ends and practical AI: speech-to-text, computer vision and local LLMs, shipped as real apps with tests and CI. I'm looking for an internship or fresher role in software development.
Portfolio · LinkedIn · Resume (PDF) · HackerRank
| Project | What it does | Built with |
|---|---|---|
| HypeSift AI | Paste a link or upload a long video, press AI Generate, get captioned, reframed shorts with titles and hashtags, and see why each clip was picked | FastAPI, faster-whisper, FFmpeg, OpenCV, Ollama, Next.js |
| Tokenomy | OpenAI-compatible gateway that cuts AI costs with exact and smart caching and cheap/strong model routing, plus a savings dashboard | FastAPI, scikit-learn, Ollama, React, Docker |
| PotSpot | Turns phone dashcam video into a map of potholes, counting each one once (live demo) | YOLO11, ByteTrack, FastAPI, React, Leaflet |
| BlackBox | Incident flight recorder: turns logs, alerts and deploys into a timeline, ranks likely triggers and drafts a postmortem | FastAPI, Drain3, React |
| TermGuard | Highlights risky clauses in rental agreements and offer letters in plain English, Hindi or Kannada, each backed by a quote | FastAPI, React, PDF.js, OCR, LLMs |
| Humanize AI | Writing suite: paraphraser with 10 modes, grammar check, summarizer and a .docx rewriter with version history | React, Node.js, Express, Gemini API |
| Qualify Milk | Predicts milk quality from seven lab readings; compares ten models honestly on a dataset where most rows are duplicates | Python, scikit-learn, Flask |
| Project | What it does | Built with |
|---|---|---|
| TaskFlow | Project and team management: Kanban board, real-time chat, role-based access, reports | Angular, Node.js, Express, MongoDB, Socket.IO |
| Beam Share | Fast, offline, verified file transfer between Android, iPhone and Windows | Flutter, Dart, BLAKE3 |
| CrowdMate | One QR code for crowded places: live virtual queues and lost-person safety tags, works offline | React, TypeScript, Node.js |
| PixelStore | Online store for PC games and gear: cart, checkout, order tracking, reviews, admin dashboard | React, Node.js, Express, MongoDB, JWT |
| Habit Tracker | Android app for daily habits, streaks and a mood journal (team of two) | Kotlin, Jetpack Compose, Room |
Also: Paroh, a hackathon team project where I worked on the front end.
Languages: JavaScript, TypeScript, Python, Kotlin, Dart, SQL
Web: React, Next.js, Angular, Node.js, Express, FastAPI, Flask, Socket.IO, Tailwind CSS
AI / ML: scikit-learn, YOLO, OpenCV, Whisper, local LLMs with Ollama, Gemini API
Data: MongoDB, MySQL, SQLite
Tools: Git and GitHub, GitHub Actions, Docker, pytest, Jupyter, Power BI
- Every project ships with a README that says what works and what doesn't.
- Tests and CI wherever the project has logic worth checking.
- Free and open tools first: most of my AI projects run fully on a laptop with no paid APIs.
- Finishing my MCA and applying for internships and fresher roles
- Building AI tools that run locally: speech, vision and language models in real apps
- Adding tests, CI and live demos to my projects
The fastest way to reach me is LinkedIn or email: pbhat554@gmail.com