I turn messy datasets into decisions worth acting on.
Graduate Computer Science (Data Science) student, Punjab Technical University. I build end-to-end analytics — from raw hospital and transaction data to dashboards, models, and automated pipelines that ship.
SELECT impact FROM anurag.projects WHERE status = 'shipped';
Projects that answer a business question, not just a technical one
Each one starts from a real dataset and ends with a number someone could act on.
Analyzed 1.2M+ transactions to quantify where inventory dollars were being wasted, and where they could be recovered.
Cleaned and analyzed the Diabetes 130-US Hospitals dataset to flag which patient segments drive readmission risk, surfaced through a Power BI dashboard.
End-to-end ML system that predicts customer churn and serves live predictions through a deployed API and dashboard — my only fully solo-built and shipped project.
An n8n workflow where four LLM agents pull trending topics, generate prompts, draft blog posts, and log results — running fully locally via Docker.
Retrieval-augmented assistant that transcribes, translates, and lets you query meeting recordings in natural language.
A voice-first booking agent built for a mock multi-location clinic — checks live availability, books appointments, sends a confirmation email, logs every booking to a sheet, and hands off to real crisis helplines if a caller is in distress.
Four-week data science internship on NHS mental health datasets — hypothesis testing, clustering, anomaly detection, and a 6-stage ETL pipeline.
What I build with
ANALYSIS
- Python (Pandas, NumPy)
- SQL / MySQL
- Hypothesis testing
- Power BI
MACHINE LEARNING
- Scikit-learn
- Clustering & anomaly detection
- Model deployment (FastAPI)
AI / AUTOMATION
- LangChain + RAG
- n8n workflows
- Vapi voice agents
- LLM prompt pipelines
DELIVERY
- Streamlit dashboards
- Docker
- Git / GitHub