About
From warehouse floors to LLM agents
I'm Joseph Rwanda — an AI engineer and full-stack architect based in Nairobi, Kenya. I build automation webapps and agent systems for operations-heavy environments.
My path to AI engineering started in enterprise operations software. After 4+ years deploying POS, WMS, and ERP systems across East Africa, I understood what production really means: connectivity that fails, labels that fade, workflows that break at scale. That operational depth now shapes how I build AI systems — with real constraints, not demo assumptions.
50+ production deployments
POS, WMS, and ERP systems deployed across manufacturing, distribution, and retail clients in Kenya and Tanzania (2020–present). Includes OrigamiWMS multi-tenant rollouts and client-specific implementations.
Selected live deployments
Production URLs and source repos for sites you can verify today.
E-commerce — fresh produce and delivery.
Company site — AIDC consulting and enterprise systems.
Photography portfolio (Canada).
This site — AI engineer positioning and case studies.
AI-assisted warehouse audit workflow (hackathon flagship).
Asset operations product surface.
Flower & herb traceability — grower, lot, and chain-of-custody workflows.
Journey
2020–2022
Enterprise Systems Foundation
Built and deployed POS, WMS, and inventory systems for manufacturing, distribution, and retail clients across Kenya and Tanzania. Learned what breaks in the field — connectivity gaps, label variance, multi-location sync.
2022–2024
Scale & White-Label
Expanded OrigamiWMS to 50+ production deployments with M-Pesa integration, offline-first architecture, and white-label multi-tenant delivery. Built Chelombai ERP (HR/Payroll/Accounting/Production modules).
2024–Present
AI Engineering Focus
Applied operational domain depth to LLM systems. Built WaybillAgent (Anthropic Opus 4.7 hackathon, top ~500 of 13,000+). Now focused on AI-powered automation webapps for operations-heavy environments.
How I work
Production over demos
Code that runs in real warehouses, not just notebooks.
Domain depth matters
AI that understands barcode variance beats generic chat wrappers.
Ship, then iterate
Staged rollouts with observable KPIs, not endless discovery.
Remote-ready, field-tested
Nairobi-based, async-friendly, US/EU timezone overlap.
Let's build something
Looking for an AI engineer who understands operations? I'm open to full-time roles, contracts, and selective consulting.