Hello, it's me

Jhon Christian
Rozano

Data Scientist

I build and experiment on machine learning and AI systems, then document what I learn along the way. This is my public repository for models, codes, experiments, and personal progress.

Illustrated portrait of Jhon Christian Rozano
What I do

Making creative and innovative solutions from complex problems.

I work where mathematics, software, and responsible AI meet.

Applied machine learning

High-impact models grounded in sound evaluation, optimization, and business context.

Enterprise AI systems

RAG, multi-agent workflows, intelligent documents, and Voice AI built for real operations.

Reliable delivery

End-to-end deployment, monitoring, governance, and cloud foundations across Microsoft Azure and AWS.

Portrait of Jhon Christian Rozano
README

Learning constantly.
Building thoughtfully.

I am a Data Scientist specializing in scalable AI solutions and machine learning models that create measurable business value. My work bridges mathematical optimization and enterprise-grade delivery.

I graduated Magna Cum Laude with a BS Mathematics degree from the University of the Philippines Diliman. Away from work, I read books, I do running, I play badminton, and I personally reflect.

Based in
Metro Manila, Philippines
Exploring
Agentic AI · AWS · Azure
Current goal
Run 200 km this year
Read the full README
Selected work

Things I build and maintain.

View all work
01

Supply Chain Inventory Optimization

Designed a Monte Carlo order-up-to policy for 90 DC–SKU pairs, modeling 26.6% less on-hand inventory while maintaining 97.9% unit fill.

Inventory OptimizationMonte CarloSafety StockBacktesting
02

Customer Segmentation

Segmented 908 customers across 42,676 transaction records into five actionable personas using Gaussian mixture modeling and PCA.

ClusteringUnsupervised LearningGMMPCA
03

Campaign Response Classification

Built a soft-voting ensemble to prioritize campaign calls, with the top 20% of ranked customers capturing 73.1% of historical responders.

ClassificationHyperparameter Tuningensembles
04

Personal Website

A static-export Next.js site for projects, data science, and coding notes deployed with Azure Static Web Apps.

2026activeNext.jsAzureportfolio
05

LeetCode Solutions Tracker

A client-side GitHub integration that summarizes solved LeetCode problems from a solutions repository.

2026activeJavaScriptGitHub APIalgorithms
My shelf

Notes.

Personal learning notes from reading data science and programming books.