Work
Projects
Projects I have actually worked on, from edtech products to ML, databases, and AI prototypes.
Overview
Named projects, specific tools, and realistic outcomes.
This list mixes product work, analytics projects, database applications, and AI prototypes. Some were shipped products, some were project builds, and some were early prototypes.
Highlights
Supplier Risk Prediction System
A machine learning project for procurement risk. The goal was to classify suppliers as low, medium, or high risk before problems show up in delivery or quality.
Cleaned supplier transaction, delivery, financial, and compliance data. Engineered features like delivery delay, defect ratio, and transaction consistency. Trained Logistic Regression, Random Forest, and XGBoost models, then built Power BI and Streamlit views around the output.
Result: Best model reached 85% accuracy. The project showed a 40% reduction in high-risk supplier selection.
View projectLumi - AI Co-Founder
A Streamlit prototype for helping someone turn a startup idea into a clearer research-backed direction.
Built a chat flow that infers the idea, asks follow-up questions, runs market and competitor research through Perplexity, and generates a structured startup report. Added PDF export so the output could be saved and shared.
Result: Working prototype for idea validation, market research, competitor analysis, and MVP planning.
iCloudSchool LMS
An edtech product built for schools and institutions under Nivati.
Worked on the learning platform, school-facing product flows, adoption, and analytics needed to make the system useful for real institutions.
Result: Adopted by 150+ schools/institutions and reached about $500K ARR.
Housing Management Database
A database application for university student housing operations.
Designed an Oracle APEX and PL/SQL application for lease tracking, rent payments, maintenance requests, occupancy reporting, and admin dashboards.
Result: Centralized housing data and reduced manual tracking across leases, payments, and work orders.
View projectOnline Grocery BOPIS Segmentation
A customer analytics project focused on Buy Online, Pick Up In-Store behavior.
Cleaned customer purchase and demographic data, used K-Means and hierarchical clustering, and visualized customer groups with heatmaps and scatter plots.
Result: Produced customer segments and recommendations for personalization, retention, and store operations.
View projectWhere I can help
For industry teams
I can help with dashboards, forecasts, internal tools, and AI workflows tied to a measurable problem.
For investors and founders
I like discussing products where AI changes cost, speed, or what a small team can realistically ship.