Llama Agents as Micro Services!!!

7,402

310 0

Published 2024-06-26

llama-agents is an async-first framework for building, iterating, and productionizing multi-agent systems, including multi-agent communication, distributed tool execution, human-in-the-loop, and more!

More details here - github.com/run-llama/llama-agents

❤️ If you want to support the channel ❤️
Support here:
Patreon - www.patreon.com/1littlecoder/
Ko-Fi - ko-fi.com/1littlecoder

🧭 Follow me on 🧭
Twitter - twitter.com/1littlecoder
Linkedin - www.linkedin.com/in/amrrs/

All Comments (20)

@1littlecoder 14 days ago

I built a 100% Local Llama Agents System - https://www.youtube.com/watch?v=aiySmi5JocQ (a bit rusty though)
@marcfruchtman9473 21 days ago

Thanks for this review of Llama Agents. This looks really impressive and useful to me. I like the idea of each agent pulling and publishing messages from a Queue. ( I am surprised they went with Orchestrator, since Architect, would have been a nice laugh). The tool plane is good to have. The fact that is using the MIT License really helps. Would be good to see this in action with local documents.
@nessrinetrabelsi8581 21 days ago

A deep comparison with langgraph would be very useful, as I was convinced that langgraph was the multi-agent framework to use for production ..
@renierdelacruz4652 21 days ago

Very well done, thanks for sharing
@rajramakalai 21 days ago

I really like the idea of it. If a self-managing or orchestrating resource/service manager and load balancer can also be added like Kubernetes, I think this will be my go-to architecture for deployment.
@rajpdus 21 days ago

I have been trying to run lambda based agents using native packaging. But needs docker for now. Orchestration of cloud functions will be fun
@srivatsa1193 21 days ago

Yeah its nice to see more distributed system engineering seep into the llm agent space. I think the right ai engineer will be good cross between a backend engineer- distributed systems + webservices/databases, somewhat of a data engineer - ETL, batch/streaming+data pipelining, LLm training and LLm ops, implementing stuff in pytorch. A background in traditional ML would be nice to have.
@KevinKreger 21 days ago

I saw Llamaindex presentation yesterday at AI Engineer World's Fair ❤
@cig_in_mouth3786 21 days ago

Perfect, still watching I hope I can use llama end to end rather than paid one, hope ollama worked
@TheRealHassan789 21 days ago

Fan of LLAMA-INDEX
@3pm479 21 days ago

I want to use it, also It would be more fun to have LLM3 agent framework.
@IdPreferNot1 21 days ago

Everything LlamaIndex releases looks like the right direction after assessing what the market needs. Does anyone know if they have deep backers to survive?
@Sanjayk0508 21 days ago

Hi, im a student, could you pls recommend some windows laptop to run models like this (to do fine tuning and stuff) and diffusion models? Thank you
@Existence05 21 days ago

I think it's quite similar to langgraph..other than monitoring ui..still thanks for sharing ❤.
@agnivamahata3870 21 days ago

make a video about agency-swarm
@harshilpatel4989 21 days ago

Kubernetes in LLM go brrrr
@yhlai0911 14 days ago

It seems to me that llama agent has not dealt with memory yet?
@jeffsteyn7174 21 days ago

😂 dammit. This keeps happening. 2 months ago I started using this architecture. Before this, I was using chatgpt to call apis. And not long after that, langchain did it. And then chatgpt itself. Ideas really are dime a dozen.
@klammer75 21 days ago

I love this! Can’t wait to play!🥳🦾😎