Hi, I’m Tonderai
I’m a software engineer with over a decade of experience, based in Berlin. I’ve spent that time building systems at companies including Klarna and On — both of which went public during my tenure. At Klarna, I’m a Senior Software Engineer on the Transaction Banking team, working on SEPA and DCL-based credit transfer and direct debit systems, merchant payout infrastructure, and real-time transaction notifications, built on AWS (SQS, Lambda) and Kafka — including leading migration efforts across these systems and mentoring engineers on design and code quality. At On, I was Technical Lead for 3rd-party logistics (3PL) and EDI integrations with Dynamics 365 ERP, leading and mentoring a team of three backend engineers — all three went on to lead teams of their own.
My work sits on the decisions that are expensive to get wrong: where service boundaries should live, how a system degrades under partial failure, what happens when an upstream payment rail misbehaves in production. The stack is mostly Node.js and TypeScript professionally; I also tinker with Rust on the side. The constant across all of it is reasoning about a system’s failure modes before they become incidents.
What this blog is about
This is where I write up the architecture and tradeoffs behind systems I’ve built — not tutorials, but the reasoning: why a design won out over its alternatives, what broke, what the failure model actually looked like, and what I’d change with hindsight. Recent examples: Handovha on keeping a signature trustworthy without accounts, Mail Koenig on building a send queue out of nothing but Postgres, and the DigitalOcean setup behind both of them — one droplet, one YAML file, no orchestrator.
If a post helped you, or you want to talk systems design, feel free to reach out.