Kaivlya N Shah

B.Tech CSE (AI & ML) @ VIT-AP '28 · Java, Python & ML · Open to internships

Hyderabad, Telangana, India

kaivlyashah@gmail.com

About

Third-year B.Tech CSE (AI & ML) student at VIT-AP University (CGPA 8.90), from Hyderabad. I build with Java and Python, from a multithreaded Java chat server to SignSense, a browser-based sign-language interpreter that runs TensorFlow Lite and MediaPipe entirely on-device (GDG Hackathon finalist). Previously Research Lead Intern at Vectr, an EdTech startup, where I built Python scraping pipelines and turned competitor data into technical requirements for the dev team. Interested in AI/ML, cloud and data engineering. Looking for internships and SDE / ML roles at product companies. Happy to chat about on-device ML, AI agents or anything you're building.

Skills

Java

Python

C/C++

JavaScript

SQL

Object-Oriented Design

Networking

REST APIs

TensorFlow Lite

MediaPipe

OpenCV

Scikit-learn

FinBERT

Express.js

React

MongoDB

Git

Linux

Data Structures & Algorithms

Multithreading

TensorFlow.js

Node.js

Google Colab

Experience

Research Lead Intern

Vectr (EdTech Startup) · Remote

Jul 2025 to Apr 2026

Built Python web scraping pipelines with BeautifulSoup to collect data across the Indian EdTech ecosystem.
Analysed competitor pricing models, user engagement strategies and content delivery approaches.
Turned findings into technical requirements for the startup's development team.

Projects

SignSense: Real-Time Sign Language Interpreter

Browser-based interpreter that turns 36 hand gestures (A-Z, 0-9) into speech, with all ML inference running on the device for privacy. Combines MediaPipe hand landmarks (21 keypoints) with a TensorFlow Lite classifier, plus a voting buffer to smooth out flickering predictions. Final round, GDG Hackathon.

JavaScript

TensorFlow Lite

MediaPipe

Multithreaded Real-Time Chat Server

Java chat server over TCP sockets using a one-thread-per-client model, handling 20+ concurrent clients. Synchronised shared state to avoid race conditions and designed a custom serialised message protocol for reliable delivery.

Java

TCP Sockets

Multithreading

FundWise: Personal Finance Dashboard

Full-stack spending tracker built at Smart India Hackathon. REST APIs connect a React frontend to a Node/Express backend, with MongoDB schemas for transaction queries and simple statistical rules that flag unusual spending spikes.

Node.js

Express.js

React

MongoDB

Education

B.Tech, Computer Science & Engineering (AI & ML)

VIT-AP University, Amaravati · CGPA 8.90/10

May 2028

Certifications

Elite + Top 2%: The Joy of Computing using Python (97%)

NPTEL

DevFestHYDERABAD

Attendee · Oct 24, 2026

About DevFest Hyderabad