Healthcare IT and data analytics. I trained as a perfusion technologist, spent over a year in the cardiac operating theatres of KIMS Sunshine and KIMS Hospitals, and worked on Hospital Information System (HIS) customization as a product intern at KareXpert Technologies. I'm now doing a PGDM in AI and Data Science in Healthcare at IIHMR Bangalore.
The projects here are me taking a question from raw data to something a clinician or a manager can read: a perfusion app built from my own clinical work, survival models, dashboards, and small apps with a tested deployment pipeline behind them.
| Project | What it does | Built with |
|---|---|---|
| IDSP disease surveillance (Telemedicon 2026) | Abstract with Renuka Prasanthi Dandi. Compared Prophet, SARIMA and linear models on 44,952 IDSP outbreak records: median forecast error rose from 77% nationally to 105% at district level, and a linear trend was never beaten. Includes a React and D3 dashboard of 16 diseases. | Python, Prophet, React, D3 |
| WholePerf | Web app for cardiac perfusion teams: BSA, blood flow, heparin and cannula calculators, per-patient procedure logs and disposables inventory, with Firestore rules that keep each clinician's records private. Live demo | React, TypeScript, Firebase |
| Chlorhexidine trial: survival analysis | Time to ventilator-associated pneumonia (VAP) across two trial arms. 103 patients and 8 events, analysed with Kaplan-Meier curves, a log-rank test (p = 0.20) and a Cox model adjusted for age, APACHE II score and gender. | Python, pandas, lifelines |
| Heart attack risk dashboard | Tableau workbook with 18 worksheets, 2 dashboards and a story on demographic, lifestyle and regional risk factors. Exploratory, not a clinical tool. | Tableau |
| Sentiment analysis dashboard | Classifies text as positive, neutral or negative (Naive Bayes, logistic regression, SVM on TF-IDF features), with a Streamlit app for typed text or CSV uploads. Demo video | Python, scikit-learn, Streamlit |
| FastAPI CI/CD pipeline | A prediction API whose tests run in GitHub Actions on every push and, when they pass, trigger a deploy to Render. | FastAPI, pytest, GitHub Actions, Render |
| Employee attrition (in progress) | Random Forest pipeline on HR data with a Streamlit front end, a FastAPI service and CI. The prediction endpoint is still being built. | scikit-learn, FastAPI, Docker |
| Crypto trader behaviour vs market sentiment | Joins a fear/greed sentiment index with trade records to compare position size and profit across sentiment phases. | Python, pandas, seaborn |
Python (pandas, scikit-learn, lifelines) · SQL and PostgreSQL · R · Power BI · Tableau · Streamlit · FastAPI · Docker · GitHub Actions
- Clinical: perfusion trainee in cardiac surgery, KIMS Sunshine Hospitals and KIMS Hospitals (2023 to 2025)
- Health IT: HIS user-interface customization for client hospitals, Product Intern at KareXpert Technologies (2026)
- Research: abstract with Renuka Prasanthi Dandi on how forecastable India's IDSP disease surveillance data is (44,952 outbreak records, 2009 to 2026), submitted to Telemedicon 2026