Full-stack developer with 5+ years building web and mobile products across fintech, insurtech, and marketplace platforms. I design scalable frontend and backend systems, then layer in AI through LLM integrations, RAG pipelines, and production-ready intelligent features.
If you're hiring for a full-stack or AI-enabled product engineering role, this page should answer the question quickly: can he architect, build, integrate AI, and ship products that move business metrics?

I'm a full-stack developer who specializes in integrating AI into real products. With 5+ years of experience, I've shipped platforms used across fintech, insurtech, and agri-commerce, from frontend UX to backend APIs and production infrastructure.
My sweet spot is the intersection of robust product engineering and practical AI. I pair OpenAI, LangChain, and vector databases with React, Next.js, Node.js, TypeScript, and Python to ship reliable, measurable features into production.
I'm on a mission to build African tech products that are globally competitive, powered by strong full-stack foundations and thoughtful AI adoption.
Full-stack and AI-enabled projects with real users, real impact, and measurable outcomes.
A production-ready, ZRA-compliant ERP platform built for small and medium businesses in Zambia. It combines invoicing, inventory, accounting, customer management, audit logs, and KPI reporting in a multi-tenant architecture with role-based access.
Zambian SMEs struggle with disconnected invoicing, inventory, and accounting workflows while still needing ZRA-compliant operations.
Built a multi-tenant architecture with role-based access, transaction-safe invoice and inventory flows, and API-first modules documented with Swagger.
Prioritized data integrity and compliance over rapid schema changes, which increases migration discipline and initial implementation complexity.
Scale by tenant isolation, index-heavy Postgres queries, cached KPI reads, and background workers for heavy reporting and reconciliation tasks.
A backend payment orchestration engine for Lenco and Stripe integrations, with subscription billing, idempotent payment processing, refund flows, HMAC-signed webhooks, and full transaction audit logs.
Payment retries, provider differences, and subscription billing can create duplicate charges and weak transaction traceability.
Used idempotency keys, a provider adapter layer for Lenco and Stripe, HMAC-signed webhooks, and full payment audit logging.
Strong auditability and idempotency introduce additional storage, queueing, and operational complexity compared to a simpler payment flow.
Run stateless API instances horizontally, process webhook and sync work in queues, and partition payment logs by tenant and time.
A full-stack fleet management platform built with a Next.js dashboard and a FastAPI backend, containerised with Docker and deployed on Vercel and Railway respectively. Manages vehicles, drivers, routes, and operations in real time โ demonstrating production-grade full-stack architecture beyond the JavaScript ecosystem.
Fleet operations data is often spread across teams, making real-time vehicle, driver, and route visibility difficult.
Separated a Next.js operations dashboard from a FastAPI service layer, with Dockerized deployments and Alembic-managed schema changes.
A polyglot stack improves flexibility but increases team context switching, tooling complexity, and debugging overhead.
Scale API and workers independently, optimize route and status queries with targeted indexing, and cache high-frequency dashboard reads.
A full-stack agri-ecommerce platform built to help small-scale farms earn more and cut post-harvest losses by connecting them directly to buyers. Features 420+ live products, real-time order tracking from checkout to delivery, and a mobile app currently in review on the Play Store.
Small-scale farms lose margin and produce due to weak market access and delayed, manual order coordination.
Built a real-time marketplace with Supabase-backed auth/data flows, payment integration, and a shared backend model for web and mobile.
Real-time updates improve user trust but increase infra cost, and synchronizing product/order logic across clients adds complexity.
Use catalog pagination, aggressive indexing for search and orders, async notifications, and staged archival for historical transaction data.
Most of my work lives inside private company repos. These highlights represent the impact and scope of production projects I've owned and shipped.View my GitHub โ
I'm actively exploring full-stack, AI-enabled, and product engineering opportunities. If you need someone who can architect platforms, build reliable systems, and integrate AI where it drives measurable value, reach out directly.