Vaishnavi Sai

At Infosys Springboard, I developed a multimodal AI application using Python, Streamlit, BLIP, and OpenAI GPT. It generated contextual stories from uploaded images using vision-language models.

I optimized the inference pipeline, reducing story generation time by 80%, and applied prompt engineering to improve coherence and contextual relevance.

For my customer churn prediction project, I analyzed and prepared customer data, then compared Logistic Regression, Random Forest, and XGBoost models. Random Forest achieved 91% accuracy.

I also implemented a federated learning framework with a three-layer defense against Byzantine attacks, integrating Kyber512 and Dilithium2 for secure model updates and communication. I'm interested in IT compliance, cloud governance, risk management, and information security, and I'm eager to learn C++ and simulation software development.