milind pius

I’m a Machine Learning Engineer focused on turning research-grade ideas into scalable, production-oriented ML systems. I build and optimize end-to-end pipelines—covering supervised learning, deep learning, NLP, and time-series forecasting—with a strong emphasis on evaluation and robustness.

In my KPMG Software Developer Intern role, I designed Python-based ML data pipelines that integrated REST APIs and structured metadata workflows, improving indexing and retrieval across 5,000+ enterprise records. I also integrated OpenAI APIs for prompt evaluation, context optimization, and response monitoring, reducing redundant API calls by 20% through structured logging and context management.

Beyond production work, I develop practical models through projects like customer churn prediction (ANN with TensorFlow/Keras and comprehensive metrics), spam detection using TF-IDF and classic NLP classifiers, and an intelligent microgrid energy management system using multivariate time-series LSTM forecasting. I’m passionate about deploying predictive systems that consistently perform in real-world engineering constraints.