Top ~3.8%
Anthropic Built with Opus hackathon (~500 / 13,000+)
Remote · Production AI · UTC+3 (Nairobi)
Production AI engineer foroperations that can't fail.
I build agentic systems with evals, MCP tool use, and enterprise integration — not chat demos. Flagship: WaybillAgent (Anthropic hackathon, top ~500 of 13,000+). 50+ POS/WMS deployments give me the domain depth to ship AI that survives real warehouses.
Hiring an AI engineer in Kenya or East Africa? Dedicated profile with local proof points and deployment context.
View AI engineer in Kenya profileFeatured work
Flagship case studies covering agents, evals, vision, and OSS — with architecture, scorecards, and outcomes.
Featured
WaybillAgent
Walk the warehouse, Claude does the audit.
WaybillAgent transforms warehouse auditing from a multi-day manual process into an AI-assisted guided walk using phone capture and agentic reconciliation—flagship build for Anthropic's Built with Opus 4.7 hackathon (selected top ~500 of 13,000+ applicants).
500 / ~13,000 (Top 3.8%)
Hackathon Cohort
2 weeks -> ~40 minutes
Audit Cycle
$2,000 scanner -> phone/glasses workflow
Device Shift
Claude Opus 4.7
Claude Managed Agents
TypeScript
Next.js
Supabase
Vercel
Featured
Outcome-Driven Agent Evaluation (Hive)
Evaluation patterns for agents that must improve real outcomes.
Exploration and extension of the Hive framework for outcome-driven agent development, focusing on how teams iterate when success is measured by business results rather than single-turn benchmarks.
Public GitHub fork
Repository
Outcome loops vs. toy task accuracy
Lens
Research and internal eval experiments
Use
Python
Agent Frameworks
Evaluation Design
OSS
Apache 2.0
Featured
AIDC Barcode Toolkit
Barcode and AIDC building blocks for Claude Code.
Open-source toolkit that packages real-world AIDC workflows so Claude Code can generate, validate, and reason about barcode and labeling tasks with domain-correct defaults.
Public OSS on GitHub
Distribution
Developer velocity for AIDC-heavy features
Focus
Domain moat meets LLM-native tooling
Bridge
JavaScript
Claude Code
Barcode Standards
Label Workflows
MIT License
Developer Tools
Top ~3.8%
Opus 4.7 hackathon (~500 / 13,000+)
50+
Production POS / WMS rollouts
3
Flagship AI case studies
OSS
Agents, evals, AIDC toolkit
4+ yrs
Enterprise & field ops
Why work with me
Production AI + eval discipline for global teams — deep GitHub activity, production rollouts, not just demos
Agents, evals & LLM systems
Claude managed agents, MCP, RAG, multi-model orchestration, and outcome-driven evals — production patterns, not demo wrappers.
Enterprise integration
POS, WMS, ERP, M-Pesa, offline-first — 50+ production deployments. AI that fits existing stacks and adoption constraints.
Domain depth & AIDC
Barcode, RFID, warehouse ops — real operational edge cases that generic AI engineers miss after go-live.
Production AI systems
Architectures and outcomes from shipped agent work — evals, integrations, and domain constraints included.
Agentic reconciliation
Vision capture → stateful walk session → ERP lookup → variance classification. Long-horizon agents with resume/retry, not one-shot API calls.
View case studyEval-driven iteration
Outcome loops and business scorecards instead of toy task accuracy. Iterate prompts, tools, and policies against operational impact.
View case studyRAG + enterprise glue
pgvector retrieval, MCP tool use, and multi-tenant Supabase — AI that plugs into existing ERP/WMS stacks with permissions and audit trails.
View case study50+
Production POS / WMS / ERP rollouts
2 wks → ~40 min
Warehouse audit cycle (WaybillAgent)
Featured in
Industry coverage on AIDC, floriculture traceability, and East Africa enterprise tech.
Skills & stack
AI-first capabilities for hire or contract—backed by operations domain depth
AI & LLM engineering
- • Claude API, managed agents, and long-horizon agent workflow design
- • MCP tooling and multi-step orchestration
- • RAG pipelines, grounding, and retrieval quality
- • Multi-model orchestration and cost optimization
- • Vision / OCR in messy real-world captures
- • Production evals and reliability patterns
Full-stack & enterprise systems
- • Next.js 14+, React 18, TypeScript
- • Supabase/PostgreSQL with RLS and pgvector
- • POS, WMS, and ERP systems with repeated production rollouts
- • M-Pesa and payment integrations
- • Multi-tenant SaaS and white-label platforms
- • Offline-first architecture for field ops
Domain expertise & ops
- • Barcode, RFID, and AIDC process literacy
- • HR, Payroll, Inventory, and Production modules
- • Warehouse operations and logistics automation
- • East African enterprise rollouts across multiple sectors
- • Remote-first delivery for global teams
Live sites
AI and OSS proof first, then enterprise and client deployments. Full case studies on the work index.
AI & OSS
WaybillAgent
AI-assisted warehouse audit workflow (hackathon flagship).
Asset Zen
Asset operations product surface with AI-assisted workflows.
Joseph Rwanda (portfolio)
This site — production AI engineer positioning and case studies.
AIDC Barcode Toolkit
OSS barcode/AIDC primitives for Claude Code and agent-assisted logistics.
Outcome Agent Evals (Hive fork)
Outcome-driven agent evaluation patterns — public GitHub fork.
Enterprise & ops
FarmTrace
Flower & herb traceability — grower, lot, and chain-of-custody workflows.
Origami Tech
Company site — AIDC consulting and enterprise systems.
Client sites
Jumbo Greens
E-commerce — fresh produce and delivery.
Shot by Mark
Photography portfolio (Canada).
Live build updates
Launches, architecture notes, and hiring signals on LinkedIn; public code and demos on GitHub—no synthetic feeds.
View LinkedInFAQ
AI engineering delivery, reliability, remote collaboration, and how engagements run
What kind of AI systems do you build?
Can you build custom Claude and MCP-based workflows?
How do you approach reliability for AI features in production?
Do you support existing products or only greenfield builds?
What does a first engagement usually look like?
How quickly can we move from prototype to production?
How does your AIDC and warehouse background help AI projects?
Are you a fit for Forward Deployed or AI Solutions Architect roles?
What timezone overlap do you offer for remote teams?
Are you open to AI engineering roles as well as consulting?
Do you work with teams outside Kenya?
Still deciding?
Email a short brief: problem, stack, timeline, and success metric—I'll reply with a realistic path.
Email JosephLet's build or hire
Open to AI engineering roles (remote or hybrid) and selective consulting for US, EU, and emerging-market teams. Share your stack, timezone overlap, timeline, and what done looks like.
Download resume
Start an AI Engineering Conversation
Share your product goals and context below. When you are done, pick one: send the inquiry straight to my inbox, or open WhatsApp with everything prefilled so you can send it there.