Master's student at IIT Patna, learning data analytics by building things instead of just watching tutorials. I code, I struggle, I break things, then I fix them, and somewhere in that loop I end up shipping something that actually works.
I like data products that go all the way: statistical anomaly detection, clean pipelines, dashboards people actually use, and a full-stack wrapper when the project needs one. Honest statistics over naive 3σ shortcuts, and things that keep working even when the AI layer doesn't.
|
Data & Analytics Full-Stack & Tools |
|
|
AI anomaly detection for any CSV or Excel file Upload a file, get auto-inferred schemas, statistical profiling, and a dual-mode anomaly engine (MAD robust-z + Tukey IQR) that resists extreme outliers. 83 passing tests, 0 accessibility violations. Next.js · TypeScript · Gemini API · Recharts |
QR-based vehicle owner contact Scan a QR sticker to reach a parked vehicle's owner without exposing phone numbers. Rate-limited plate verification blocks remote abuse, and the whole thing runs at zero cost. Supabase (RLS) · Next.js · React · Vercel |
Multi-year sales analytics in Power BI Interactive regional sales views with DAX measures isolating profit margin trends and top customer cohorts, built to inform inventory and distribution decisions. Power BI · DAX · SQL · Excel |
- Data Analytics Virtual Internship, Deloitte (Forage), Jun 2025
- Python for Data Science, NPTEL (IIT Madras), Jul to Aug 2024
- Database Management System, NPTEL (IIT Madras), Jul to Sep 2025
Open to Data Analyst, BI, and entry-level Data Science roles


