Deep dives into systems architecture, AI patterns, and the future of work. We share what we learn from the trenches of software engineering.
AI
How to build a production-grade RAG pipeline using Pinecone, Hybrid Search, and Re-ranking. We discuss the trade-offs between dense and sparse vectors and how to optimize for latency.
Agents
Decoupling high-cost reasoning from low-cost execution. A technical look at how we implemented self-healing scrapers that use LLMs only to 'plan' the route, reducing scraping costs by 90%.
Vision
Why traditional OCR fails on complex tables. We benchmark IBM's Docling against Tesseract and Azure AI Document Intelligence, showing how 'Vision-Language Models' are changing the game for data extraction.