-
How I spent a weekend reinventing a 50-cent chip
An optocoupler is a part I’d seen in schematics for a while and quietly skipped over. There’s an LED pointed at a photo-sensitive transistor with a dashed line between them. The dashed line is the entire point: the two halves of the part are not electrically connected. The signal crosses as light, over a few hundred microns
-
When speed becomes strategy 💨
Startups thrive on speed and hustle. But when everything feels urgent, how can you tell what actually matters?
-
A pint with Jimothy: on fear, ego, and hiring smarter
☕️ We started with coffee, like most chats do. Black for me, cappuccino for Jimothy (he actually prefers Jim, or James). Corner terrace, Singelgracht. One of those confusing Amsterdam afternoons where the sky can’t decide − sunlight breaks through the clouds, then ducks back behind them. Just long enough to let a few leaves blow
-
When Slack starts to feel like a DDoS attack
In software engineering, we often rely on “exponential back-off” when retrying failed network requests − a technique where each subsequent attempt is spaced out further in time to avoid overloading the system. Oddly enough, I’ve found myself applying a similar concept to human communication. As an Engineering lead, I’m frequently on the receiving end of
-
AI for Engineering managers: adapt now or trail behind
Remember when a five‑digit Stack Overflow score was a flex? Today that, and a vintage 2022 playbook will buy you precisely zero leverage. Yesterday’s job, tomorrow’s irrelevance Many Engineering managers still run on three rituals: However, none of those moves the product faster. Meanwhile, AI agents are quietly doing code reviews, generating boilerplate, even writing RFCs. The org chart hasn’t
-
DIY Bluetooth Kobo page-turner remote with ESP32 and Lilka
If you’ve ever read an e-book with your hands tucked under a cozy blanket: apparently there are page-turner remotes just for that. And in this post, we’ll DIY it with an ESP32-based device called Lilka.
-
A minimal LLM Ops stack with tracing and model costs
I built a minimal FastAPI “customer support reply drafter” with TF-IDF retrieval and Langfuse tracing. You’ll see exactly what context the model used, where latency came from, and what each request cost, plus the trade-offs behind the design.
-
RAG: A (mostly) no-buzzword explanation
Retrieval-Augmented Generation (RAG) is a pattern that fixes the knowledge cutoff and hallucination problems by giving an LLM access to the right data at answer time. Instead of asking the model to “remember everything”, RAG lets it look things up first, then answer.
-
Unikernels, without the marketing
It all started when I saw Prisma put “serverless” and “no cold starts” in the same sentence describing their “Prisma Postgres” product 🤔
Looking for more posts?
For contact options, head to About and Contact