Notes on building AI software that doesn't fall over.
Practical posts on agents, LLM pipelines, FastAPI, AWS, and the trade-offs that come with each. Subscribe via RSS.
From Streamlit Prototype to Production: A Checklist
Streamlit is great for prototypes and bad for production. Here's how to know when to keep it, when to wrap it, and when to rewrite.
- #streamlit
- #python
- #production
- #deployment
Hiring an AI Agent Developer in 2026: What to Look For
A practical checklist for founders and engineering managers hiring their first AI agent developer — beyond the buzzwords.
- #hiring
- #ai-agents
- #engineering-management
Deploying FastAPI on AWS Lambda — A Minimal, Real-World Setup
The shortest path from a working FastAPI service to a public HTTPS endpoint on Lambda, with cold start mitigations and real cost numbers.
- #fastapi
- #aws
- #lambda
- #python
- #deployment
Extracting Structured Data from PDFs with Tesseract and LangChain
A two-stage pipeline that turns scanned invoices, contracts, and forms into typed JSON your back office can actually use.
- #ocr
- #tesseract
- #langchain
How I Build Production AI Agents with LangGraph
A practical breakdown of the architecture I use to ship reliable LangChain/LangGraph agents — typed state, deterministic nodes, observability, and human-in-the-loop.
- #langgraph
- #langchain
- #ai-agents
- #python