Engineering Insights

Demystifying the 'B' in LLMs—understanding how model size affects performance, reasoning, and local deployment.

Implementing structured output enforcement, auto-correction loops, recursive category discovery, and metadata-first extraction for a resilient AI agent.

How to handle large documents in local LLMs using Map-Reduce summarization, token counting, and production-grade configuration.

Building a smart, local-first file organizer that understands document context and automates classification using Go streaming pipelines and Apple MLX.

How I resolved performance bottlenecks and stabilized a high-traffic Node.js service early in my career by implementing MongoDB connection pooling.