About me
I build data and ML systems that stay fast, reliable, and cheap to run.
Over the last few years, I’ve built real-time payment layers, event pipelines, and automated workflows across fintech, gaming, and healthcare—routinely halving latency and cutting cloud bills by over a third.
I prefer working directly with users to dig through messy requirements, stress-test assumptions, and turn real-world edge cases into clean, reliable features shipped straight to production.
What I'm Doing
-
Payment Integration and APIs
Integrating Stripe and M-Pesa into high-availability Go payment systems with Redis caching, idempotent webhooks, and sub-second validation.
-
Enterprise Systems — ERP, POS and E-Commerce
Architecting resilient backends for ERP operations, POS terminal integrations, and e-commerce platforms with automated multi-channel inventory synchronization.
-
Data Pipelines and Event Streaming
Building distributed streaming pipelines with Apache Kafka, PySpark, AWS Glue, and Airflow processing millions of high-velocity daily events.
-
Production ML and Predictive Systems
Shipping real-time churn, lifetime value (LTV), and spatial valuation models with Python, FastAPI, PyTorch, scikit-learn, and Docker.
-
AI Integration and Intelligent Workflows
Embedding modern LLM APIs, retrieval-augmented generation (RAG), and agentic task automation into operational business tools and products.
-
Cloud Infrastructure and Cost Optimization
Cutting cloud spend by 35%+ via containerized AWS Lambda and Docker architectures with robust PostgreSQL audit trails and resource rightsizing.
Professional Impact
-
Labs Afrika
Partnered directly with merchants to clarify messy edge cases and transaction flows, integrating Stripe and M-Pesa into a high-availability Go payment layer. Cut validation and serving latency by over 50% while reducing cloud infrastructure spending by over 35%.
-
Rapid Research
Built high-throughput ETL pipelines with AWS Glue and PySpark, transforming millions of daily gaming events into structured feature sets for model evals. Shipped real-time churn and lifetime value models delivering high-freshness predictive signals to operational teams.
-
TechFlow Solutions
Automated healthcare data reconciliation in Python, saving dozens of engineering hours weekly and eliminating manual spreadsheet errors. Migrated hundreds of gigabytes of clinical data to AWS S3 with strict schema validation.
-
GitHub Campus Expert
Led interactive hands-on workshops on Git, open-source contribution practices, and CI/CD pipelines, mentoring developer communities on modern software engineering standards.
