CS senior, IIIT, RGUKT, Ongole. Data Analyst / Data Engineer track.
nagamohan.me · LinkedIn · X · nagamohan765@gmail.com
I pick a real dataset, a real API, or a real institution's data, and I build something that has to work against messy, undocumented, inconsistent reality — because that's the only way you find out you don't understand something yet.
status = {
"data_engineering": "pipelines, storage formats, performance — in progress",
"full_stack": "React frontends on FastAPI/Node backends",
"analytics": "dashboards that answer a question, not decorate a report",
"applied_ml_nlp": "models that solve something, not demo something",
}
RaktRadar — live blood-shortage tracker on India's eRaktKosh government API. The government's own district codes don't match the API's internal IDs. Found that out the hard way, then rebuilt every lookup table directly from eRaktKosh's DOM and AJAX endpoints instead of trusting the published code sheets. That's the whole project in one sentence: government data lies about its own schema, and you build around it.
TRACE — Transit Analytics & Reliability Correlation Engine, on real Delhi DTC bus data. GTFS static feed plus a live Protobuf real-time feed. Batch-first ETL, star schema (trip schedule, trip actual, complaints, weather), built to answer reliability questions a transit authority would actually ask. DMRC data excluded on purpose — scope stays honest.
UDISE Education Analytics Pipeline — Python ETL into PostgreSQL, modeled through dbt (staging to marts), visualized in Power BI, with a Gemini API layer on top generating insights from the marts instead of raw numbers.
Built a grievance management portal running in production across 15+ police stations, and a hostel management system used by 1,500+ students. FastAPI, ReactJS, MySQL. Both still running. Neither was a prototype.
Python · SQL · FastAPI · ReactJS · PostgreSQL · MySQL · dbt · Power BI · PySpark · Airflow · Streamlit · Azure