AI Engineering in Practice11 min read
User and session models for multi-tenant AI agents
How to model users and conversation sessions for multi-tenant AI agents. The schema, the tenant guard, and the pattern that scales to thousands.
Loading...
Build Your First Software Factory Execution Harness live Thursday, Oct 1, 12:00 PM ET
Oct 1, Reserve a seatBlogDaily bytes of
Daily bytes of
AI engineering.
Bite-sized insights for building production AI systems. Expert guides, real-world patterns, and practical engineering wisdom.
Explore posts
136 posts in total
AI Engineering in Practice10 min read
Connection pooling in production Python AI services
Why your agentic AI service hits 'broken pipe' errors at 100 users and how SQLAlchemy connection pools fix it. The sizing rule and the lifespan pattern.
AI Engineering in Practice10 min read
SQLModel for agentic AI: beyond JSON blob storage
How SQLModel gives agentic AI systems type-safe structured storage. The end of JSON blob columns, the Pydantic bridge, and the 50-line pattern.
LLM Engineering11 min read
Hybrid retrieval in RAG: vector + graph search
How hybrid retrieval combines vector search and graph traversal in RAG. The when, the why, and the 60-line fusion that beats either alone.
AI Engineering11 min read
bash tools for coding agents: safe shell execution
How coding agents run shell commands safely. The bash tool design, the denylist trap, and the sandbox that keeps rm -rf away from your home directory.
LLM Engineering11 min read
RAGAS evaluation for RAG pipelines: a practical guide
How to use RAGAS to evaluate RAG pipelines. The 4 metrics that matter, the eval loop, and the trap that makes most RAG evals dishonest.
LLM Engineering12 min read
Agentic RAG with LangGraph: planning, rewriting, tool use
How agentic RAG uses LangGraph to plan, rewrite queries, and call tools. The graph, the state, and the nodes that turn single-pass RAG into reasoning.
LLM Engineering12 min read
Dynamic RAG: re-planning retrieval strategies mid-pipeline
How dynamic RAG re-plans retrieval mid-pipeline. The replanner prompt, the loop that recovers from bad chunks, and when this beats single-pass RAG.
LLM Engineering11 min read
JSON output parsing for RAG: grounding with Pydantic
How JSON output parsing grounds RAG answers in retrieved context. The Pydantic schema, the prompt, and the validation that kills hallucinations.
LLM Engineering11 min read
LangChain chain types: stuff vs map reduce vs refine
Stuff, Map Reduce, or Refine? How to pick the right LangChain chain type for RAG summarization, and the cost and quality trade-offs that decide it.
LLM Engineering11 min read
Advanced RAG: quote extraction for context compression
How quote extraction shrinks RAG context by 80% without losing answer quality. The pattern, the prompt, and the code that ships in production pipelines.
AI Engineering12 min read
Build a coding agent with Claude: a step-by-step guide
How to build a coding agent with Claude in 100 lines of Python. Tool use, file editing, bash, and the loop that ties them together. No frameworks.
Your Instructor
Built by Param Harrison
Cofounder of AEOsome.com and Chief Mentor at learnwithparam.com with 14+ years building production systems. I've trained 50+ engineers on AI engineering - these programs distill what actually works into structured paths you can follow at your own pace.
Ready to go deeper?
Go beyond articles. Build production AI systems with hands-on courses, or join a live cohort.