I’m currently learning Machine Learning, Reinforcement Learning, RAGs and LLM applications .
- Reach me at: Linkedin
- Portfolio: Shubham's Portfolio
I’m currently learning Machine Learning, Reinforcement Learning, RAGs and LLM applications .
Production-grade clinical trial success/failure prediction: calibrated probabilities, conformal intervals, SHAP explainability, drift-triggered retraining, and a scoped K8s deploy.
Jupyter Notebook
A domain-agnostic ELT warehouse platform - pharma regulatory data as the reference implementation.
Jupyter Notebook
Transformer encoder built from scratch in PyTorch for biomedical NER (BC5CDR) without HF models. Includes a fine-tuned BERT baseline comparison, tests.
Python
Global Sales Analytics is an interactive app using Python, SQL, and Streamlit to explore sales data across regions, countries, customers, and products, enabling analysis of revenue trends, order pa…
Python
Predicted customer churn using UCI Online Retail dataset. Designed 3NF SQLite database, engineered RFM features. Conducted 16 experiments with 4 algorithms using Optuna tuning,tracked via MLflow an…
Python