Projects

Research and machine learning projects by Pujari Keerthika. For peer-reviewed work, see Publications.

Predictive Modelling in Lung Cancer

Deep learning applied to medical imaging for early-stage lung cancer detection. The project combines predictive analytics with AI-driven prognosis, with the goal of catching malignancies earlier than conventional screening workflows and supporting better patient outcomes. It reflects a broader research interest in medical imaging — one of the areas where machine learning can have the most direct human impact — and builds on the same computer-vision foundations as her published 3D reconstruction work.

PFAS Source Tracing Using Machine Learning

View on GitHub

Machine learning classification of PFAS ("forever chemicals") contamination sources in environmental datasets. The approach uses fingerprint analysis across fish tissue samples to attribute contamination to its likely origin — turning raw environmental chemistry measurements into actionable source identification. The project sits at the intersection of ML and environmental science, and grew out of Keerthika's research exposure to computational chemistry during her internship at IIT Madras.

NLP for Intent-Based Networking (Kubernetes)

View on GitHub

Fine-tuned large language models on Kubernetes documentation to automate intent-based network configuration. Instead of hand-writing configuration, an operator expresses what the network should do in natural language, and the model translates that intent into concrete configuration — enabling AI-driven network management with minimal human intervention. This project came out of her research internship work on intent-based networking at IIT Madras and combines her NLP interests with practical systems engineering.