Dr. Maryam Miradi

Пікірлер

@RedCloudServices 4 күн бұрын

Don’t forget Pydantic AI! 😊

@maryammiradi 4 күн бұрын

Definitely not! Will make a video on that one! Anything Pydantic will last. In fact most of these frameworks are in reality Pydantic with their own twist.

@JoramMillenaar 7 күн бұрын

Hey Dr. Maryam! Thank you for taking the time to share your insights. I think you might have missed an important part of the prompt tuning process; training. You mentioned the random state of the embedding, which is initially true, but the power comes in when you train that embedding.

@maryammiradi 7 күн бұрын

Thanks alot! You are absolutely right I didn’t focus on Training in this video. It was more to show how prompt tunning and prefix are different than fine tunning and inspire people to try it.

@ArpitGupta-g4x 9 күн бұрын

Ma'am can u please re-upload your time series' project

@maryammiradi 7 күн бұрын

Not in this channel anymore as it is dedicated to AI Agents perhaps in another channel or a mini-Training.

@ArpitGupta-g4x 7 күн бұрын

@@maryammiradi can we please except some kick ass ai agents project here , for free ! Please look after the economically back followers of urs

@nigamlenka3957 10 күн бұрын

Data science guide

@maryammiradi 7 күн бұрын

Of course! Please go ahead and get it here: www.maryammiradi.com/free-guide

@dineshnallabothula6723 10 күн бұрын

Hi Maryam, I’ve been following you on LinkedIn and truly appreciate the insightful knowledge you share. I wanted to learn more about the AI Agents Mastery program. Does it include real-world projects with all the associated challenges, or are the datasets simplified for learning purposes? Additionally, do participants work on the complete end-to-end project development lifecycle, from development to production deployment? Are best practices for building and deploying agentic solutions covered in detail? Does the program also explore LLM architecture, including how to fine-tune open-source models for real-world use cases? For example, do we learn how to adapt models like Falcon LLM or LLaMA for specific applications? Does it focus on using open-source models or paid ones? Moreover, does the program cover the evolution of agentic systems, such as transitioning from simpler automation to multi-agent systems or autonomous AI agents? Are metrics for evaluating agentic systems, such as performance, collaboration efficiency, or decision-making accuracy, discussed in-depth? Lastly, is there an emphasis on designing these solutions using microservice architecture versus monolithic architecture? Thank you!

@razdingz 12 күн бұрын

beneficial informations

@maryammiradi 12 күн бұрын

Glad to hear!!

@jaggyjut 13 күн бұрын

something i'm still confused is when to use langchain, crewai or llamaindex.are there specific use case for when to use one framework or combine them together? Appreciate the guidance.

@maryammiradi 3 күн бұрын

Here’s one way to think about it: LangChain is the most general “LLM plumbing” library out there-it helps you chain together prompts, handle agent logic, manage memory, parse outputs, and integrate with lots of vector stores, APIs, and models. If your use case involves building multi-step workflows, adding “memory,” or having an LLM act on your behalf (agents), LangChain is a great fit. LlamaIndex (formerly “GPT Index”) focuses on indexing and retrieving data for LLMs. It’s excellent if you need to feed large amounts of custom or private data to a language model in an efficient way. It supports multiple ways to structure and query your data (e.g., tree indices, vector indices). Often you’ll see LlamaIndex combined with LangChain-LlamaIndex handles the data ingestion/retrieval side, and LangChain handles the rest of the LLM workflow (prompt chaining, agent logic, etc.). CrewAI (sometimes spelled “Crew AI”) is a bit newer and less widespread than LangChain, but it aims to do some of the same orchestration and workflow-building around LLMs. It’s intended to be more user-friendly and collaborative, focusing on how you’d design your prompts and chain them together in a more visual or team-based environment. If you prefer a simpler, more guided approach than LangChain, CrewAI could be a good solution. So in practice: If your main need is a robust library of ready-made “building blocks” (chains, memory, tools, agents), LangChain is typically the go-to. If your biggest challenge is hooking up lots of domain-specific content to an LLM for retrieval-based Q&A, LlamaIndex is a strong choice. If you want a lighter-weight, possibly more visual framework (but still want to piece together multi-step LLM workflows), CrewAI might be enough on its own. You can absolutely combine them. One common pattern is: Use LlamaIndex to manage and query your private data efficiently. Plug those queries (retrieval steps) into LangChain so that the model can use that data in a more advanced chain/agent workflow. If you need a more visual or collaborative approach (e.g., a low-code environment for your team), you could fold in CrewAI for the “front-end” of your prompt design and orchestration.

@jaggyjut 3 күн бұрын

@@maryammiradi Thank you so much for taking the time to respond to my query and for explaining it so clearly! Your detailed explanation really helped me understand the concept better, and I truly appreciate the effort you put into it. Keep up the amazing work-you’re making a real difference!” 🙏

@maryammiradi 2 күн бұрын

@@jaggyjut Really glad to hear it is useful! and Thank you for the kind words!!

@somnathdhn12345 18 күн бұрын

Free data science guide

@maryammiradi 17 күн бұрын

Of course! Go ahead and grab it here: www.maryammiradi.com/free-guide

@saaj-l9n 19 күн бұрын

Free data science guide

@maryammiradi 17 күн бұрын

Of course! Please find your free guide here: www.maryammiradi.com/free-guide

@authentic-reflections 2 күн бұрын

Free data science guide

@CHRISTYLARA-k1h 19 күн бұрын

LLM made me laugh, but now it’s making me think. Launching on SOLANA and trading on BingX tomorrow… Could this be more than just a meme? I’m ready to find out!"

@bulentoz2810 25 күн бұрын

Free Data Science Guide

@maryammiradi 25 күн бұрын

Sure! Go ahead and Grab the Data Science Guide here: www.maryammiradi.com/free-guide

@RedCloudServices 29 күн бұрын

Dr. Miradi can you make a video with this use case which replaces the RAG pipeline with a Vision Language Model which can understand images, text, charts and tables directly?

@maryammiradi 28 күн бұрын

Of course! I can do that in my future videos

@RedCloudServices 28 күн бұрын

@ ❤️

@CortezLabs Ай бұрын

Thank you

@alwikah5664 Ай бұрын

Free data science guide

@maryammiradi Ай бұрын

Of course! Go ahead and download it here: www.maryammiradi.com/free-guide

@abhimane__ Ай бұрын

Can u suggest text to speech or speech to text conversational. Without paid api key.

@maryammiradi Ай бұрын

Yes, of course! In this video, I show one which is free and on Hugging Face: Hands-On AI Agents: 12X BEST tips with LangGraph, LangChain, CrewAI, OpenAI Swarm & Hugging Face LLM kzbin.info/www/bejne/r4LRgGN7ZqqEpKs

@ouskah2469 Ай бұрын

Free Data Science Guide

@maryammiradi Ай бұрын

Of course! Here you can reach the free guide: www.maryammiradi.com/free-guide

@jaggyjut Ай бұрын

Awesome. Does the langchain/langgraph training include concepts like RAG, Real-Time Audio, Vision & Multi-Modal Models or part of the Complete package?

@maryammiradi Ай бұрын

Yes, it includes RAG, Audio (Text to Speech), Vision models from Hugging face. Check out the program on Training page: www.maryammiradi.com/ai-agents-mastery

@orlandoagostinho Ай бұрын

Thanks

@AdilButt-s8j Ай бұрын

❤❤❤

@techievans Ай бұрын

Free Data science Guide

@maryammiradi Ай бұрын

Of course! Please go ahead and grab it here: www.maryammiradi.com/free-guide

@srikanthdongala6922 Ай бұрын

Free Data Science Guide

@maryammiradi Ай бұрын

Of course! Go ahead and grab it here: www.maryammiradi.com/free-guide and you can explore my website here: www.maryammiradi.com

@errrr3 Ай бұрын

Data science guide :)

@maryammiradi Ай бұрын

Of course! Please find the free guide here: www.maryammiradi.com/free-guide and you can explore my website here: www.maryammiradi.com

@ahmedaq9018 2 ай бұрын

Thanks a lot Maryam!!!!!! ❤❤❤❤

@IsxaaqAcademy 2 ай бұрын

Very helpful thank you But can you start from the basics: 1. Create a new dataset 2. Push it to hf 3. Finetuning for non-english

@maryammiradi 2 ай бұрын

Sure! I will try to make a video on that.

@gft01 2 ай бұрын

Sorry. Maar war was dat over het Nederlands. Werkt u in Nederland? Of woont u er?

@maryammiradi 2 ай бұрын

Ik ben idd een Nederlander, woon en werk in Nederland

@gft01 2 ай бұрын

@maryammiradi mooi toch. Weet u de huidige stand van ai crackdetection of soortgelijke processen? ..toevallig? Gaan de ontwikkelingen snel of juist niet