Multi-Agent & Applied AI Deep Dive

30 minutes, screen-shared, real work. We'll dig into:

• Multi-agent architecture — when to use LangGraph vs. simple tool-use loops, state design, error recovery • Memory layers — mem0, vector stores, when not to use them

• Async workers & queues — Taskiq / Celery / SQS patterns for LLM calls that take 30+ seconds

• Edge AI / IoT — model selection, MQTT pipelines, low-power inference

• "Build vs buy" decisions for AI features

What you'll leave with:

A clear next step you can ship this week Specific patterns/repos to copy from A written summary I send within 24 hours

Best for senior engineers, tech leads, and founders shipping AI features to real users.