scroll
Denzel Chingodza
"Humour: 75 percent. Honesty: 90 percent."
— TARS, Interstellar
About
I am a software engineer working across backend systems, cloud infrastructure and applied AI. I build in Python first, define infrastructure as code, and care about systems that are reliable, observable and understood, not just shipped. I understand the problem before I write code, which usually means talking to the people who have it. Using AI tools to move faster is a habit now, and I love the research that comes before and alongside the build. But the goal is always to understand what I build, not outsource the thinking. One day I want to build my own AI company, one that's more careful about where it sits in law and society (haha, we'll get there).
Outside the code
I write on Medium about building with AI. First piece: The Unlicensed Builder. Serious basketball player. My faith in Jesus Christ is what's utmost important to who I am and what I do. It's the anchor of the balance of my life. I also lead a small group of men at church, where we help each other balance life, careers and opportunities through that same lens. And a New York Knicks fan. Not a casual one. First chip since 1973. Brunson Finals MVP. I built an entire app about it because apparently that is a normal thing to do when your team wins.
Currently
Software Engineering Fellow at Funema. Funema is a venture capital firm that invests in and co-builds technology with founders across emerging markets. I work across backend and AI on a live, AI-powered product: mapping how its services, APIs and data fit together, improving the stability and quality of the codebase, reviewing AI-assisted code with care, and documenting changes for future development and handover.
Building into
Full-stack systems end to end: clean APIs, solid databases, maintainable frontend, reliable infrastructure. On top of that, NLP, RAG pipelines, and LLM fine-tuning (BLOOMZ-3B for Shona is live). I care about the engineering, not just the output. Every tool I use, I try to understand what's actually happening under the hood.
Location
South Africa. Open to remote work and relocation for the right opportunity.
Writing
The Unlicensed Builder
On building real systems with AI, the gap between output and understanding, and why the craft still matters.
Read on Medium ↗What Do We Make of AI, in the Age of AI?
A personal resolution on faith, technology, and what it means to build in this moment as a Christian believer.
Read on Medium ↗Work
Serverless URL monitoring on AWS with an AI incident digest.
EventBridge triggers Lambda checks every 60 seconds. Every resource defined in Terraform, reproducible from one command. An SQS dead-letter queue ensures no check is silently dropped. The AI layer reads 7 days of monitoring data and returns a plain-English summary.
A Shona-language conversational AI fine-tuned from BLOOMZ-3B.
Fine-tuned with QLoRA on free compute only. Found and fixed an 84% data contamination issue in the training set. Built a 15-prompt, 7-category evaluation scorecard because no standard Shona benchmark exists.
RAG document Q&A built for medical students at Stellenbosch University's Tygerberg campus.
Students spent up to 45 minutes searching a 200-page PDF for a single clinical criterion. Now they get answers grounded in the document, with page-level citations.
Routes plain-English questions to the right service and returns the answer.
A natural language interface that sits on top of my own projects. You ask a question, it decides which system to call, and returns the result. One endpoint to rule them all.
Built to understand the cross-section of business and software — where financial metrics live in the database.
Models the full subscription lifecycle — MRR, ARR, churn, LTV, ARPU, NRR — with real SQL, 100 seeded customers, a FastAPI layer, and a live dashboard. Built to understand how the numbers a business runs on actually live in a database.
Explore all projects
Click a planet
Skills
Languages
Backend and APIs
Cloud and DevOps
Data and AI
Databases
Contact
If you want to work together, talk through an idea, or have an opportunity you think I'd be interested in, send a message. I actually check my email.