Shreya Mane

I'm a computer science undergraduate at IIIT Tiruchirappalli building at the intersection of machine learning and business intelligence. I delivered a 50,000+ record customer churn pipeline that achieved 89% accuracy and supported retention decisions through interactive dashboards.

As a Data & Machine Learning Intern at I Studio, I developed a face-recognition pipeline using PCA Eigenfaces and a backpropagation neural network. I benchmarked seven configurations, identified k=50 as the strongest stable setting, and implemented imposter rejection with a 0% false acceptance rate on non-enrolled held-out faces.

I also architected AutoDAN-Evo, a black-box evolutionary framework for evaluating LLM alignment robustness. It generated human-readable jailbreak suffixes and achieved a 50–70% jailbreak success rate on Vicuna-7B using adaptive mutation and multi-objective fitness scoring.

I work across Python, SQL, Scikit-learn, Power BI, and data visualization to clean data, build models, and communicate findings that matter.