B.Tech Data Science student turning raw data into decisions. I craft interactive Power BI dashboards, run end-to-end EDA, and ship machine learning models — deployed live on Hugging Face Spaces.
I'm a final-year B.Tech Data Science student at Moradabad Institute of Technology, passionate about the full data lifecycle — from cleaning messy datasets to deploying machine learning models that solve real problems.
My core strength is data analytics with Power BI, Python & SQL, backed by hands-on machine learning across NLP and recommendation systems. I've completed simulations with TATA and PwC, building dashboards and optimizing ETL pipelines.
I love shipping work the world can actually use — that's why my ML projects are live on Hugging Face Spaces.
Analyzed and cleaned 100K+ records using SQL and Power Query. Built interactive Power BI dashboards, performed EDA, and generated business insights using DAX and visualization techniques.
Built interactive dashboards with drill-through reports, optimized ETL pipelines, and developed KPI-monitoring dashboards using Power BI and SQL.
Completed training in Python, Machine Learning, Data Analysis, Deep Learning, and real-world project implementation.
NLP-based spam classifier using TF-IDF and Naive Bayes, achieving 95%+ accuracy. Type any message and get an instant Spam / Not-Spam prediction. Deployed on Hugging Face Spaces.
Content-based recommendation engine using cosine similarity. Pick a movie and get the top similar titles with posters. Built with Streamlit and deployed on Hugging Face.
Interactive Power BI dashboard analyzing 9,600+ titles across countries, ratings, genres, and release years using DAX, Power Query, drill-through reports, and rich visual storytelling.
Advanced deep learning web application that detects and classifies 5 eye conditions (Cataract, Glaucoma, Myopia, Diabetic Retinopathy, Healthy Eye) with 96.69% accuracy. Uses EfficientNetV2 and CLAHE local contrast enhancement. Deployed on Hugging Face Spaces.
AI-powered, fully browser-based facial recognition attendance system. Uses face-api.js for real-time face detection, registration from 5 angles, auto-marking, and a glassmorphic dashboard with date-wise records and CSV export. No backend needed.
Local MVP for marketing performance reconciliation and optimization. Features PDF invoice extraction using a custom Python parsing engine, dynamic data tables, SQLite/PostgreSQL storage, session-managed auditor login, and a Vercel-ready serverless backend.
Have a dataset that needs decoding or a model that needs building? I'm always open to interesting problems and collaborations.
✉ sharmasharad792@gmail.com