Shubham Shirodkar

My published, peer-reviewed research on maternal health risk classification involved auditing and revising the pipeline and building a Voting Ensemble classifier. I also delivered a Tableau Public dashboard and reviewer-response document for the publication process.

For the SOLiGence Barometer Ensemble System, I designed a multi-model forecasting platform combining BiLSTM, XGBoost, TCN, and TFT-Lite models. I added a LightGBM meta-learner, a regime-detection gate, and a FinBERT-based sentiment pipeline, then shipped a FastAPI and Streamlit dashboard with GitHub Actions CI/CD.

I'm building STRIDE, a sensorless, markerless screening system for identifying movement patterns linked to injury risk. I validated pose extraction against lab-grade OpenCap data and developed an uncertainty-weighted screening layer and counterfactual explainability module.

I also built a customer churn classification pipeline covering feature engineering, model selection, and evaluation. I'm completing an MSc in Artificial Intelligence and Data Science at Solent University and am seeking entry-level Data Science or Data Analyst roles in the UK.