Posts by David Park — Harbor Software

LangChain vs Building Your Own: When Frameworks Help and When They Hurt

LangChain has become the default framework for building LLM applications. It has 65,000+ GitHub stars, extensive documentation, and integration with every LLM provider and vector database on the market. It is also one of the most controversial tools in the AI engineering community, with vocal

ETL vs ELT: Making the Right Choice for Your Data Stack

The ETL versus ELT debate has been running for over a decade, and the conventional wisdom has shifted dramatically during that time. Five years ago, the default recommendation was almost always ETL: extract data from sources, transform it in a dedicated processing layer (typically Spark

Redis Beyond Caching: Queues, Pub/Sub, and Rate Limiting

Redis is one of the most misunderstood tools in the modern software stack. Most teams discover it as a cache — a faster alternative to hitting the database for frequently accessed data. They install it, configure a TTL, and move on. That is a valid

Document Parsing with AI: Extracting Structure from Chaos

Businesses run on documents. Invoices, contracts, purchase orders, spec sheets, compliance filings — the operational backbone of most companies is a sprawling mess of PDFs, Word files, scanned images, and emails with attachments. The information locked inside these documents is critical for operations, compliance, and

Database Migrations in Production: A Zero-Downtime Approach

It’s 3 PM on a Tuesday. Your migration is halfway through altering a 200-million-row table. The application is throwing 500 errors. Your CEO just messaged the engineering Slack channel with a screenshot of a blank page. The database is locked. You can’t roll back because

Stripe Integration Patterns for Subscription SaaS

Why Stripe Integration Is Harder Than It Looks Every SaaS founder hits the same wall. You sign up for Stripe, drop in Checkout, and think billing is solved. Three months later you’re debugging webhook race conditions at 2 AM while a customer’s subscription is stuck

API Design Patterns That Scale

Every API starts simple. A handful of endpoints, a few hundred requests per day, one or two client applications. Then it grows. New consumers appear — a mobile app, a partner integration, an internal analytics dashboard. Request volume goes from hundreds to hundreds of thousands.

Kubernetes in Production: What They Don’t Tell You

Every Kubernetes tutorial ends at the same place: you have a deployment, a service, and an ingress. Your application is running. kubectl get pods shows healthy green status. The tutorial declares success and moves on. What the tutorial does not cover is the next twelve

Building Resilient Microservices with RabbitMQ

Synchronous HTTP calls between microservices are a reliability liability. When Service A calls Service B over HTTP and Service B is slow or down, Service A is also slow or down. The caller inherits the callee’s failure mode. Multiply this by ten services and you

Author: David Park