Health informatics · Clinical analytics · New York
I'm Rutul — a physician-trained health informaticist. I build the audit tools, validation workflows, and dashboards that make sure what coders, providers, and AI systems say about patients actually holds up.
ScrollEvery decision in healthcare — clinical, operational, financial — runs on data. I make sure that data is accurate — especially now that AI is doing the reading.
The stack
Analysts don't know what the data meant at the bedside. Physicians don't build the dashboards. Coders don't see the whole system. The overlap is the point.
01 / CLINICAL
Thirty patients a day. Fifty-plus critical procedures. A clinical internship through a pandemic. I've been the person writing the chart — so I know what a diagnosis looks like when it's real, and what it looks like when it's paperwork.
02 / INFORMATICS
SQL, Python, Tableau, Power BI — plus the clinical intelligence layer: CDSS design, EHR implementation, population health analytics. The training to turn messy clinical reality into something legible.
03 / OPERATIONS
Coding audits, data reconciliation, KPI reporting, and project management across a growing clinical network — promoted within a year. I run the data operations that keep a value-based organization honest with itself. That's the job.
Selected work
Real problems from a real healthcare operation — anonymized, but every outcome is measured.
Risk-adjustment audits used to depend on what each reviewer happened to remember. I built a dashboard that maps every HCC code to explicit confirmation criteria and cross-references provider and census data — turning tribal knowledge into a repeatable, defensible process.
Admission-discharge-transfer feeds, consult logs, and facility censuses all disagreed — and nobody could say which was right. I built matching pipelines that reconcile all three and flag discrepancies, including outside trips that never hit the census.
Avoidable hospitalizations are expensive for value-based contracts and worse for patients. I built the risk-identification and care-coordination workflow that connects data flags to actual provider follow-up across teams.
Provider performance conversations ran on anecdote while the data sat unused. I built monthly KPI dashboards and provider ranking tools across facilities, designed for a non-technical leadership audience.
A multi-facility expansion into New Jersey with no dedicated project manager. I coordinate the rollout: staffing trackers, program materials, and clinical program proposals including tele-cardiology diagnostics.
Provider schedules lived in inconsistent spreadsheets that had to be manually reconciled every cycle. I wrote a reusable Python + SQLite pipeline that ingests, normalizes, and queries schedule data on demand.
The lab
Side experiments — same instinct as the day job: take something messy and make it legible.
A public, browsable HCC reference on GitHub Pages — the lookup tool I wished existed while doing validation work.
A personal practice tracker with a contribution-style heatmap. Deployed and in daily use.
A Scriptable home-screen widget with live fixtures and results for the tournament.
Scrapes deal subreddits and surfaces matches straight to the iOS home screen.
54 ASCCP-guideline care pathways with 93% recommendation accuracy — graduate work that previews the validation mindset.
Regression on clinical trial data predicting outcomes and length of stay; feature engineering lifted accuracy 18%.
Drag / scroll →
Writing & research