Senior Cloud & DevOps Architect & AI/LLM Engineer • 13+ Years Experience

I design and scale cloud-native infrastructures across AWS, GCP, and Kubernetes, and I build production LLM / GenAI applications on top of them. Most of what I ship now is full-stack product work: Next.js on Vercel, Postgres, and Claude doing the job that used to need a back office.

valupro · CRM and valuation report software for Kenyan registered valuers. A firm runs enquiries and quotes, inspects and values a property, passes QA checks, then issues a signed, sealed PDF report that banks verify by QR code. Next.js 16 and React 19, Neon Postgres with Drizzle, Vercel Blob for photos and seals, Paystack billing in KES by card or M-Pesa, PDFs via @react-pdf/renderer.

Transcribez · AI transcription and translation SaaS built on Anthropic's Claude. Live in production with 900+ users. Prompt engineering, structured segment output, multi-language translation, anti-hallucination guardrails, and model routing (Sonnet/Haiku) for quality versus cost.

mpesa-mock · Local M-Pesa Daraja API emulator. Run STK push, C2B, B2C and callbacks on your own machine instead of fighting the sandbox. Published to npm and GHCR.

npx mpesa-mock

Cloud Architecture    →  Multi-cloud design on AWS & GCP, cost optimization, security
DevOps & Platform     →  Terraform, GitOps, CI/CD pipelines, zero-downtime deployments
Kubernetes            →  Production clusters (EKS/GKE), Helm, operators, service mesh
Data Engineering      →  Databricks, Spark ETL, Airflow, streaming pipelines
AI/ML Engineering     →  Inference APIs, MLOps, TensorFlow, PyTorch
LLM / GenAI           →  Production apps on Claude, prompt engineering, structured
                         outputs, RAG, agents/tool-use, evals, cost & latency tuning
Product Engineering   →  Next.js on Vercel, Postgres/Drizzle, auth, billing, PDF
                         pipelines, M-Pesa and Paystack

Cloud & Infrastructure

Languages & Frameworks

Data & ML

LLM / GenAI & Product

• AI Automation · Enterprise customer engagement and agentic (MCP tool-use) workflows
• Data Platform · Databricks lakehouse with streaming ETL pipelines
• Observability · OpenTelemetry collectors with multi-backend routing
• Platform Engineering · Self-service developer tooling and GitOps workflows

FinTech • HealthTech • PropTech • EdTech • Cybersecurity • Telecom • Government & NGO • Logistics • E-commerce

_{Open to consulting, architecture reviews, and interesting infrastructure challenges. Reach out at [email protected]}