In our pipeline, a FastAPI endpoint receives the request. In one request, different agents can run — 2 from the Orchestrator (planning + combining), 4 from the ExcelAgent, 1 from the Researcher (which handles RAG and web search), and 1 from the Image Generator. That means there can be up to 8 LLM calls in total. Right now, I’m using concurrent.futures to run the agents in parallel, and everything runs on Celery. Is this the best approach, or is there a more standard/efficient way to handle this concurrency?

I’m building an AI system and trying to understand when to use a workflow (deterministic pipeline where I hardcode the steps) versus an agent (LLM-driven, decides which tools or actions to take at runtime). From what I’ve read, things like prompt chaining or routing are considered workflows, while tool-using agents are more autonomous.

I’m currently learning how to build AI agents, and while I already have a gist of the main frameworks and even tried them myself, I still feel a bit overwhelmed about which one is best suited for production. When building a production-level AI agent, which approach do you prefer — LangGraph, AutoGen, CrewAI, OpenAI’s Agents SDK, or just raw LLM calls? I’d love to hear your perspective on why you’d choose one over the others.

Today i had an experience in my business and it reminded me that success isn’t a straight road, it’s full of detours, slow days, and unexpected wins. funny how Some mornings I feel unstoppable, and others I wonder if I’m moving at all hahaha. Well guess what, every struggle has shaped me, every small win has fueled me, and every lesson is part of the bigger picture. Honestly balancing life and business isn’t perfect, but I’m learning daily that progress matters more than perfection. If you’re in a tough spot, HOLD ON. If you’re in a good one, CELEBRATE IT. Either way, trust me, you’re growing. What’s one thing this journey has taught you about yourself?

👉 https://chatgpt.com/g/g-sroxms7qG-python-tudor 👈 Wow guys I have something for you. I built This old project a while back when GPT Agents first came along. And Ironically I think this is perfect for learning Python in this community really for free. I plan on reviving it tuning under the hood with API, and Python, and putting a wrapper to overlay as an app. Deffinently make it independent of ChatGPT. Let me know what you guys think of this project. Give some insight on how to improve the transition.