Everything here is a public repository. Each one opens on GitHub, research pipelines, applied AI work, and the systems coursework underneath it.
github.com/kinjalumassPlugging PII leakage in the chain-of-thought of large reasoning models
A reasoning model can answer carefully and still give the private data away in its working. The chain-of-thought is treated as scratch space, so it leaks the personal information the final answer was written to protect. This work sanitizes the reasoning trace itself, without breaking the reasoning that depends on it.
kinjalumass/BCU-Graph-Analysis
Data science for safer and more equitable biking infrastructure in Boston, built with the Boston Cyclists Union through the 2026 Data Science for the Common Good program. Graph analysis over the city road and cycling network, turning urban mobility and road-safety data into evidence that holds up in a policy conversation. Engineered as a reproducible project rather than a notebook dump: a DVC pipeline, a pytest suite, GitHub Actions running tests on every pull request, and Sphinx documentation.
kinjalumass/CS690FStartup
Built for COMPSCI 690F, Trustworthy and Responsible AI. A synthetic image-and-caption dataset of ten classes, a baseline classifier trained on it, and membership inference attacks run against that classifier to test what it leaks about its own training data. The same question as her published research, approached from the attack side.
kinjalumass/kinjalumass.github.io
Angular 22, built from scratch. Three sections with separate design languages, lazy-loaded per route, and an assistant grounded in a corpus generated from the same data that renders the pages, running on Cloudflare Workers AI.
kinjalumass/Project-Report-3-602-Phenomena-Exploration
Research methods coursework for COMPSCI 602: exploring a phenomenon empirically and writing it up to the standard the field expects.
kinjalumass/CODSOFT
The project set built during the CodSoft AI internship, covering natural language processing, computer vision, and neural networks, each taking an algorithm from paper to working code.
CryptoProj1
Implementation work from MATH 4175, Cryptography at Virginia Tech: classical and symmetric-key ciphers built from the primitives rather than called from a library.
Crypto-Proj2
The second cryptography project from MATH 4175, continuing into cryptanalysis and public-key constructions.
BombLab
The classic defusing exercise: reading x86 assembly and reverse-engineering each phase to work out the input it wants.
understanding-assembly-code
Working through compiled output to understand what the compiler actually produces.
unscramble
Recovering structure from jumbled regular and binary data.
c-practice
Small C programs written while learning systems programming.
3114-proj2
Coursework for CS 3114, Data Structures and Algorithms at Virginia Tech.
Proj3
Coursework for CS 3114, Data Structures and Algorithms at Virginia Tech.
website
An earlier site build in TypeScript.