Latest from the build floor
Most LLM agents follow the same pattern: read the user’s query, call some tools, gather data, and respond with a data-aware answer. It works […]
// Recent posts
How a routine autovacuum on a high-traffic Postgres table flipped a query from 1.3ms to 5 seconds and pegged our Aurora reader at 100% CPU. A walkthrough.A quiet Monday morning. A little too quiet for our liking. The…
After a year of wrestling the monstrous reconciliation beast, building fortresses, taming wild data, and navigating the chaos of automation, it’s time to look ahead.Reconciliation isn’t a solved problem. It’s a living, breathing challenge that grows as we scale,…
Part 5: The Human Element — How Reconciliation Shapes Creator Experience and FinanceIntroduction: Why Humans Matter in ReconciliationBy this point in our reconciliation journey, we’ve wrestled with data dragons, tamed monstrous files, and built complex state machines to keep order statuses…
Part 4: Automated Reconciliation — The Dream, The Reality, and The Lessons LearnedBy now, we’d built a solid fortress: manual tangles were tamed, data was chunked and sanitised, servers stopped melting, and order statuses marched in line. The next natural step?…
Part 3: The Order Status State Machine — Making Sense of the MadnessBy now, we’d built a fortress that could handle uploads, chunk huge files, and keep servers alive. But reconciliation isn’t just about data volume — it’s about meaningful data. And that…
We’re growing at a breakneck speed at Wishlink. We’re onboarding 100s of creators every month, over 1m users shop from us on a daily basis, and we capture millions of clicks and sales every month. All of this…
Part 2: Building the Fortress — From Manual Mishaps to Automated WorkflowsWe had just pulled back the curtain on the monstrous beast that was our reconciliation process — bloated data, tight timelines, edge cases, and chaos at every corner. Now comes the…
// How we work
Start from lived problems, then work backward to the system.
Prefer small, observable bets over perfect plans.
Design today’s primitives for tomorrow’s scale.
