Self-taught, shipping in production.

I’m a Google Cloud Certified ML Engineer and backend developer. I learned by building and shipping systems, not through a bootcamp. Today I run Axene.io, a B2B infrastructure platform giving African SMEs local-first, M-Pesa-compatible alternatives to USD-priced SaaS. Email, meetings and a CRM are live today; chat, payments, domains, analytics and ERP tools are next on the roadmap.

I’ve mentored rising data professionals, led AI/ML programming for GDG Chuka and She Code Africa, and spend a lot of weekends in rooms full of laptops helping someone ship their first model. I care about the whole lifecycle, from data engineering to training to containerised deployment on GCP.

Things I’ve built.

Axene.io: business infrastructure

End-to-end engineering of Axene Mail, Axene Meet and an integrated CRM, built as affordable alternatives to USD-priced SaaS for the Kenyan SME market.

120 customers on Axene Mail

FastAPI, PostgreSQL, Docker, GCP, SMTP/DNS

Nia: gamified CBC revision

An AI tutor named Doma guides students through the CBC curriculum, with separate student, teacher and parent roles.

Winner, Claude Hackathon, Elimu track

React 19, TypeScript, Node/Express, Neon Postgres

Duka Store: M-Pesa-ready online shop

A Django storefront for everyday Kenyan goods, with category browsing, a cart and checkout built around M-Pesa payments, currently running against the M-Pesa sandbox.

Django, M-Pesa API

Document Q&A with RAG

Ask questions about your own documents and get answers grounded in the source text, using retrieval-augmented generation.

RAG, LLM

Website for C&W Advocates

2026, an Axene.io client project

A fast, mobile-friendly site for a Nairobi law firm, launched on the firm’s own domain and organised around what a client needs to decide: practice areas, partner profiles and how the firm works.

Experiments and notebooks

Smaller ML projects, run on synthetic or sample data to test an idea.

River: streaming ML pipeline

A pipeline that imputes missing values, scales and one-hot encodes features, then learns a logistic regression one row at a time, scored by predicting before it learns. Run on a deliberately messy synthetic Kenyan customer dataset.

Python, River, online ML, logistic regression

Impute & Validate: KNN vs mode

Fills missing values in a Kenyan customer-transactions dataset, then tests the result by hiding known values and scoring the guesses against a mean/mode baseline.

KNN beat the baseline: MAE about 653 vs 775

Python, Scikit-learn, KNNImputer

Neural Networks & Agents Lab

October 2026

Three runnable demos on synthetic data: a PyTorch credit-risk network, a random-forest check on which features drive a lending decision, and a rule-based script that logs maintenance tickets for failing rural water pumps.

Python, PyTorch, scikit-learn

Where the work has lived.

Founder & AI/Python Lead

Axene.io
  • Architected and shipped Axene Mail, a transactional email platform on FastAPI, PostgreSQL, Docker and GCP that replaces USD-priced services for Kenyan SMEs, now with 120 customers.
  • Built AI features inside Axene Meet (bookings) and the CRM. Chat, payments, domains, analytics and ERP modules are planned, not yet shipped.
  • Led the full SDLC on consulting projects, from system design to containerised deployment.

Data Science Mentor

Data Ladder
  • Mentored aspiring data scientists through hands-on cohorts covering Python, statistics and applied ML.
  • Guided mentees from raw datasets to deployed models, reviewing notebooks, feature engineering and evaluation choices.
  • Introduced standardised visualisation frameworks so cohorts could read and present their results consistently.

Lead Data Analyst

Mentor Me Collective
  • Led collaborative data pipelines in Pandas and BigQuery for cohort analysis at scale.
  • Built Scikit-learn models to predict student placement.

AI/ML Community Lead (volunteer)

She Code Africa & Google Developer Groups, Chuka
  • Designed and delivered AI/ML workshops and hackathons using TensorFlow, PyTorch and Scikit-learn.
  • Ran GCP workshops on Vertex AI, Cloud Functions and serverless AI architecture.
  • Organised community Buildathons that produced working cloud-native AI prototypes.

Talks and community.

Panel at the Africa Technology Show 2026: four speakers seated on stage in front of a screen showing the panel title, with Beth Njeri Kimani on the right
Africa Technology Show 202623 July 2026. Panel on digital public infrastructure and youth-led startups
RenderCon Kenya 2026 talk
RenderCon Kenya 202617 October 2026. Building AI features real SMEs will actually use
THX The Hotel Expo at KICC Nairobi
THX: The Hotel Expo 202621 August 2026, KICC, Nairobi
Women in Tech Summit Kenya talk
Women in Tech Summit KenyaUpcoming, 27 November 2026. From prototype to production
Talk at GDG Chuka
GDG ChukaThe Intelligence Revolution: how smart systems are redefining modern software
Generative AI talk at She Code Africa Chuka
She Code Africa ChukaGenerative AI talk
Also
She Code Africa × PinkHub AI/ML sessions. AWS Student Builder Group at Zetech University: Intro to SQL with Amazon Athena.
Mentoring
Data Ladder, Cohort 6 graduation.
Awards
GoMyCode Kenya Hackathon, KES 300,000 scholarship.

What I build with.

Languages
Python, SQL, TypeScript
ML and AI
PyTorch, TensorFlow, Scikit-learn, XGBoost, River, LangChain
Backend
FastAPI, Django, Node/Express, React
Infrastructure
GCP (Vertex AI, Cloud Run), AWS, Docker, CI/CD, SMTP, DNS
Data
PostgreSQL, BigQuery, Pandas
Tooling
Git/GitHub, PyTest

Building for African SMEs, or hiring for an engineering role?

I’m open to work right now: backend, AI and ML roles are what I’m after, and I’ll keep building Axene alongside. If you’re hiring, or have a problem an African SME would pay to have solved, write to me.

njeri@axene.io