Building machine learning models, data pipelines, and automation systems, then deploying them so people can actually use them.
[ Machine Learning ] [ Data Engineering ] [ Backend APIs ] [ Automation ]
📍 Nagpur, India
Machine Learning & Data Science
AI Automation & Workflows
Backend & Databases
Analytics & Interface
Infrastructure & DevOps
- ◈ Agentic Workflows: Learning how to build multi-agent systems that run tasks on their own and call tools.
- ◈ RAG Systems: Exploring retrieval-based LLM apps to make answers more accurate.
- ◈ MLOps: Learning how to deploy, monitor, and update ML models in production.
- ◈ Production ML APIs: Building FastAPI services that serve models in real time.
01 · Fraud Detection System
An end-to-end system that scores transaction risk in real time.
| Stack | Python · LightGBM · FastAPI · Streamlit · Render |
| Features | 434 behavioral and identity features |
| Result | 0.9405 AUC with LightGBM (best of 4 models tested) |
| Repository | View Repository |
- Built the features from raw transaction data.
- Compared 4 models and picked LightGBM.
- Served the model through a FastAPI backend.
- Built a Streamlit page for interactive predictions.
- Deployed on Render.
02 · Olist E-Commerce SQL Analysis
SQL analysis of a large Brazilian e-commerce dataset.
| Stack | MySQL · SQL · CTEs · Window Functions |
| Scale | 100,000+ orders across 9 tables |
| Key finding | Orders delivered 3+ days late saw review scores drop from 3.29 to 2.10 |
| Repository | View Repository |
- Designed the schema with primary and foreign keys.
- Built an ER diagram.
- Used multi-table JOINs, CTEs, and window functions.
- Created a seller health score from revenue, delivery speed, and reviews.
03 · Los Angeles Crime Intelligence Dashboard
A 5-page Power BI dashboard built for non-technical decision makers.
| Stack | Python · SQL · Power BI |
| Scale | 1M+ LAPD crime records over 6 years |
| Key findings | 9% arrest rate · 12-day average reporting delay |
| Repository | View Repository |
- Cleaned and processed over 1 million rows.
- Wrote plain-language KPIs so anyone can read the dashboard.
04 · Automation Systems
LinkedIn → WhatsApp Automation
Scrapes LinkedIn posts, picks relevant ones, and sends them to WhatsApp automatically.
Python · Web Scraping · WhatsApp Automation · Render · View Repository
Faculty Intelligence Scraper
Crawls college websites and pulls faculty details into Excel and Google Sheets.
Python · Web Scraping · Excel · Google Sheets · View Repository
Automation & AI Intern · Incubein Foundation Nagpur, Maharashtra · June 2026 – Present
- ◈ Built a LinkedIn scraping bot that sends curated posts to WhatsApp, deployed on Render.
- ◈ Built a faculty scraper that collects names, contacts, and departments into Excel and Google Sheets.
- ◈ Built outreach automation workflows for the incubator.
- ◈ Shipped a live website on Supabase, Netlify, and GoDaddy using OpenCode and Lovable.
- ◈ Research Publication: Real-Time Energy Monitoring Using Energy Meter (IJSRED / Meerayan Journal) · DOI:
10.5281/zenodo.15226430 - ◈ What I built: Added WiFi remote start, stop, and reset to a smart energy meter that only monitored before.
- ◈ Data Science & Analytics with Artificial Intelligence · NSDC / Skill India · 2026
- ◈ IBM Certified: Data Analysis with Python · 2025
- ◈ Google AI Professional · 2026
- ◈ HackerRank SQL Intermediate · 2025
Network: Portfolio / LinkedIn / Email / GitHub
"Build it, ship it, measure it, improve it."


