Real-world learnings from building and operating lending and debt collection platforms. This pillar documents what actually works in production, what breaks under pressure, and the patterns that emerge when you build these systems at scale.
Captures the messy reality of building financial systems. These aren't polished case studies but active documentation of what we're learning as we build. You'll find notes on production incidents, architectural decisions that worked (and ones that didn't), and the kind of operational knowledge that only comes from running systems in the real world.
6 topics
These notes document how missed payments are detected late when payment workflows, transaction outcomes, and account entity updates fail to converge in time, and how that delay shapes pre-collections actions, intent signals, and escalation behavior.
When payment outcomes are detected later than they occur, systems can act on outdated assumptions. This can affect how pre-collections actions are timed and interpreted by customers. Treating detection timing as an explicit part of the system helps keep responses aligned with what is actually known at the time.
Modeling debt collection lifecycles as state machines. From account and data onboarding through pre-arrears, early warning, arrears onset, handling transitions, edge cases, and state-dependent business rules.
Modeling loan lifecycles as state machines. From origination through payoff, handling transitions, edge cases, and state-dependent business rules.
Understanding idempotency, retry logic, and reconciliation patterns when handling payments at scale in lending platforms.
Building automated collections workflows that balance recovery rates with customer experience. Dunning strategies, payment arrangements, and escalation paths.
Check out topics from other pillars or browse all topics.