01Live product
AI platform / Africa
AI-Sprouts
A platform making AI tools and training more available across Africa. I built the data pipelines, model integrations, and APIs that hold the product together.
- Next.js
- Express
- TypeScript
- AI integration
Lawrence Wafula / backend engineer
I build the parts that keep products moving.
I design APIs, data models, auth flows, and deployment systems for teams that need their software to stay understandable after it ships.
2021started building with Python
04live products
KEbased in Kenya
01 / Selected work
Work that has to hold up after launch.
A few products I have helped build, ship, and keep running. The common thread is useful software with a clear path from idea to production.
02Live product
Infrastructure / security
ChamaAI
Infrastructure and security work for a product using Supabase for data and auth, custom DNS, and a deployment path that does not need a keyboard nearby.
- Supabase
- DNS
- Security
- Deployment
03Live product
Autonomous GTM platform
Greywolfx Leads
An autonomous GTM platform built on Convex and Clerk. It finds and enriches leads, scores them, then runs the outreach campaigns end to end.
- Convex
- Clerk
- Lead scoring
- Outreach automation
04Live publication
Technical publication
Runtime Signals
A technical publication about building reliable AI-agent systems in production. I run the editorial pipeline and the security, performance, and accessibility gates the site has to clear before anything ships.
- Astro
- TypeScript
- Editorial workflow
- Security
02 / About
The work is mostly invisible. That is the point.
I'm Lawrence, a backend engineer in Kenya. I work across APIs, data models, authentication, and deployment so the visible parts of a product have something solid underneath.
I like the practical middle ground: enough structure to make change safe, not so much ceremony that the system becomes its own problem.
Make failure boring
Error handling, retries, logs, and useful defaults are part of the product, not cleanup for later.
Keep the shape clear
Small modules, honest names, and documentation that helps the next person make a change safely.
Treat security as structure
Auth, secrets, permissions, and deployment boundaries belong in the first version of the system.
A short timeline
2021
Started with Python
Small scripts and CLI tools turned into a lasting interest in how software behaves under pressure.
2022
Services, not just scripts
FastAPI, Django, REST, and the project structure decisions that start to matter once other people depend on the code.
2023
Put it on the internet
Docker, DNS, Supabase, security hardening, and the practical work of keeping public software healthy.
Now
Making useful systems calmer
Building products end to end while keeping the backend legible, resilient, and ready for the next change.
Currently splitting time between product work, infrastructure, and learning where useful separation ends and needless complexity starts.
03 / Stack
Tools are useful. Judgement is the stack.
These are the tools I reach for most often: enough to ship a product, diagnose a problem, and leave the system better than I found it.
Languages
PythonTypeScriptJavaScriptSQLBash
Frameworks
FastAPIDjangoExpressNestJSNext.jsReact
Infrastructure
DockerPostgreSQLSupabaseRedisDNS
Practices
REST APIsAuth flowsSecurityCI/CDMicroservices
01
Designing boundaries
Turning product requirements into APIs and data models that are easy to reason about.
02
Making releases repeatable
Containerised deployments, sane configuration, and fewer steps that depend on memory.
03
Connecting the whole product
Working closely with the frontend so the contract feels coherent from request to screen.