Notes & Experiments

Thoughts on product, engineering, AI, and infrastructure.

POSHardwareLatencyVenue OperationsOffline FirstPerformance

Hardware Constraints and Real-World Latency in Venue POS Deployments

The gap between lab benchmarks and what actually runs on venue handhelds during a rush is brutal—here’s what we’ve measured in production and the design choices forced by real hardware.

POSReconciliationConfigurationSnapshottingFinancial AccuracyVenue Operations

Configuration Snapshotting for Accurate Financial Reconciliation in High-Volume Venues

Price changes, tax rules, and discount logic that shift mid-event can silently corrupt end-of-night totals—snapshotting every lane’s effective configuration at transaction time is the only reliable way we found to make reconciliation match reality.

POSEvent-DrivenInventoryCRDTPartition ToleranceStadium Operations

Event-Driven Inventory Management for Stadium POS: Consistency Across Partitioned Nodes

Keeping beer, hot dogs, and merch counts sane across 80+ lanes while half the venue loses connectivity for 15 minutes requires more than optimistic locking—here’s the event-sourced model that actually survived real games.

POSOffline FirstIdempotencyQueueingNetwork PartitionVenue Operations

Offline-First Transaction Queuing in Venue POS Systems: Idempotency Under Network Partitions

When the network dies mid-rush, the only transactions that survive are the ones you can prove happened exactly once—here’s how we built queuing and idempotency that actually works under real venue partitions.

POSOffline FirstBurst LoadVenue OperationsEvent-DrivenResilience

POS Systems for High-Volume Venues: Lessons Learned from Burst Loads and Offline Resilience

Running point-of-sale systems in stadiums and large venues taught me that burst capacity matters far more than steady-state throughput, and graceful degradation during connectivity loss is non-negotiable.

AIProduct ManagementAutomation

How AI Agents Can Replace 30% of Product Management Meetings

Exploring how AI agents can automate status updates, decision logging, and alignment checks — cutting meeting overhead by nearly a third.

DevOpsSelf-HostingInfrastructure

Optimizing Traefik for Multi-Host Self-Hosting

A practical guide to configuring Traefik as a reverse proxy across multiple Docker hosts with automatic TLS and dynamic routing.

Product ManagementPOSArchitecture

POS Systems for High-Volume Venues: Lessons Learned

What I learned building point-of-sale systems for stadiums and arenas where speed, reliability, and offline-first are not optional.