Hey, I'm

Jesse Ngugi

A Data scientist & analyst at LuxDev. Turning raw data into decisions, and code into products.

I build end-to-end data pipelines, craft predictive models, and design dashboards that actually get used. Starting my computer science degree this September — excited to deepen the theory behind the work I already love.

View projects Get in touch

About me

The story so far, and where I'm headed.

I'm a data scientist and analyst currently working at LuxDev, where I spend my days cleaning messy datasets, building predictive models, and translating numbers into narratives that drive business decisions.

My toolkit spans Python, SQL, and the usual suspects in the data stack — but what I really enjoy is the problem-solving: figuring out why a model underperforms, or designing an experiment that actually isolates causality.

This September I'm starting a computer science degree, which I'm treating as a chance to go deeper on the theory I currently use intuitively. I want to understand algorithms from first principles, not just import them from scikit learn.

When I'm not coding, you'll find me exploring new datasets for fun, writing about what I learn, or debating whether pandas is actually better than polars.

Languages & core tools

Python SQL R JavaScript Git

Data & ML

Pandas scikit-learn TensorFlow PyTorch Spark Tableau Power BI

Engineering

Docker AWS FastAPI PostgreSQL MongoDB

2025 — 2026

Data scientist & analyst

LuxDevHQ

September 2026 — September 2030

BSc Computer Science

Cooperative University

Future

Data scientist / software engineer

Building what's next

Quick facts

Based in Nairobi, Kenya
Fluent in English, Swahili [+ a bit of Spanish]
Open to freelance & collaborations
Always learning something new

Selected projects

A few things I've built. Short descriptions, real impact.

Customer churn predictor

An end-to-end machine learning system that predicts customer churn risk and explains why each prediction was made. Built on the IBM Telco Customer Churn dataset, the model uses XGBoost with a tuned decision threshold to hit 80%+ precision, and every prediction comes with a plain-language explanation generated from SHAP values — not just a probability score. Served via a FastAPI endpoint and fully containerized with Docker.

Python XGBoost FastAPI Docker SHAP scikit-learn

Real-time sales dashboard

An ETL pipeline that cleans and normalizes retail sales data (9,994 orders) into a properly structured PostgreSQL database, then feeds a two-page Power BI dashboard. Rather than loading data as one flat table, the schema is normalized into customers, products, orders, and order line items — enforcing real referential integrity. A simulation script demonstrates live data ingestion by inserting new synthetic orders into the database on demand or on a schedule, visible in a dedicated "Live Feed" dashboard page.

Python ETL PostgreSQL Power BI SQLAlchemy

Social sentiment analyzer

NLP pipeline that classifies brand mentions across Twitter and Reddit into sentiment buckets. Fine-tuned DistilBERT with custom preprocessing. Achieved 91% accuracy on test set.

PyTorch HuggingFace NLP spaCy

Automated data pipeline

Scheduled Airflow DAGs that ingest, clean, and validate data from 6 external APIs. Includes data quality checks with Great Expectations. Saves ~15 hours of manual work per week.

Apache Airflow Python AWS S3 dbt

Blog

Notes on data, code, and the things I learn along the way.

Aug 2026 6 min read

Why I stopped using pandas for everything

Polars, DuckDB, and the rise of query engines that make pandas feel like a relic. A practical migration guide with real benchmarks on a 10M row dataset.

Jul 2026 8 min read

Building a churn model that actually gets used

Most ML models die in a Jupyter notebook. Here's how I turned a prototype into a production system that the sales team checks every morning.

Jun 2026 5 min read

SQL window functions I wish I knew earlier

A no-fluff guide to LAG, LEAD, ROW_NUMBER, and the patterns that make complex analytical queries readable and fast.

Curriculum vitae

Download my full résumé or view the highlights below.

Jesse Ngugi — CV

Data scientist & analyst with experience in predictive modeling, data science, and business intelligence. Currently at LuxDev, starting CS degree in September 2026.

Download

Get in touch

Open to collaborations, freelance work, or just a chat about data.

Email

ngugigichuki21@gmail.com

Location

Nairobi, Kenya

GitHub

github.com/ngugiauraa

linkedin.com/in/ngugiauraa