Armand Ruiz

HUGE NEWS - Last big AI update of the year from our friends at Meta. Welcome Llama 3.3! This is all you need to know about it:

𝟭/ 𝗛𝗶𝗴𝗵 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗮𝘁 𝗥𝗲𝗱𝘂𝗰𝗲𝗱 𝗦𝗰𝗮𝗹𝗲: Llama 3.3 is a 70B parameter model that delivers the performance of a 405B parameter model, making it significantly more efficient.

𝟮/ 𝗖𝗼𝘀𝘁-𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆: The model is easier and cheaper to run, enabling wider accessibility.

𝟯/ 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗱 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗧𝗲𝗰𝗵𝗻𝗶𝗾𝘂𝗲𝘀: Leverages the latest post-training innovations, including online preference optimization, to enhance performance.

𝟰/ 𝗢𝗽𝗲𝗻-𝗦𝗼𝘂𝗿𝗰𝗲 𝗔𝗰𝗰𝗲𝘀𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆: Designed to make state-of-the-art AI more accessible to the open-source community.

𝟱/ 𝗜𝗺𝗽𝗿𝗼𝘃𝗲𝗱 𝗖𝗼𝗿𝗲 𝗖𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝗶𝗲𝘀: Achieves better results while maintaining cost-effectiveness, benefiting developers and researchers.

Oh, and Next stop -- Llama 4 in 2025 🚀

- Download from Meta: https://lnkd.in/gPK9QzxM
- Download on Hugging Face: https://lnkd.in/grD-eV2B
- Available TODAY in watsonx.ai
- IBM Press Release: https://lnkd.in/ghiNZ_JS

Forget RAG, welcome Agentic RAG

𝗡𝗮𝘁𝗶𝘃𝗲 𝗥𝗔𝗚
In Native RAG, the most common implementation nowadays, the user query is processed through a pipeline that includes retrieval, reranking, synthesis, and generation of a response.

This process leverages retrieval and generation-based methods to provide accurate and contextually relevant answers.

𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗥𝗔𝗚
Agentic RAG is an advanced, agent-based approach to question answering over multiple documents in a coordinated manner. It involves comparing different documents, summarizing specific documents, or comparing various summaries.

Agentic RAG is a flexible framework that supports complex tasks requiring planning, multi-step reasoning, tool use, and learning over time.

𝗞𝗲𝘆 𝗖𝗼𝗺𝗽𝗼𝗻𝗲𝗻𝘁𝘀 𝗮𝗻𝗱 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲
- Document Agents: Each document is assigned a dedicated agent capable of answering questions and summarizing within its own document.

- Meta-Agent: A top-level agent manages all the document agents, orchestrating their interactions and integrating their outputs to generate a coherent and comprehensive response.

𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀 𝗮𝗻𝗱 𝗕𝗲𝗻𝗲𝗳𝗶𝘁𝘀
- Autonomy: Agents act independently to retrieve, process, and generate information.

- Adaptability: The system can adjust strategies based on new data and changing contexts.

- Proactivity: Agents can anticipate needs and take preemptive actions to achieve goals.
Applications

Agentic RAG is particularly useful in scenarios requiring thorough and nuanced information processing and decision-making.

A few days ago, I discussed how the future of AI lies in AI Agents. RAG is currently the most popular use case, and with an agentic architecture, you will supercharge RAG!

Stop trying to hire the smartest people for AI roles.

In my previous role, we hired 475+ people.

For sure we looked for good AI technical knowledge, but to truly stand out, we highly valued:

- Get stuff done
- Avoid drama
- Trust each other
- Laugh together
- Communicates well
- Wants to learn and grow
- Work as a team

I always look for those qualities more, just like Chris Donnelly suggests.
___

If you like this content and it resonates, follow me Armand Ruiz for more like it.

Have 10 hours? IBM will train you in AI fundamentals - for free

IBM's AI Fundamentals program is built inside of our SkillsBuild learning portal.

The credential includes six courses.

1/ Introduction to Artificial Intelligence
2/ Natural Language Processing and Computer Vision
3/ Machine Learning and Deep Learning
4/ Run AI Models with IBM Watson Studio
5/ AI Ethics
6/ Your Future in AI: The Job Landscape

Optional:

7/ Mastering the Art of Prompting

You can get started here: https://lnkd.in/gu5Rtu-n

P.S. Are you an IBMer? Repost this ♻️ for your network, so we get more people to learn AI with us. Thank you!

RAG vs. Fine-tuning: It's not an either/or choice

The debate around whether Retrieval Augmented Generation (RAG) or fine-tuning yields better results for LLMs often misses the point.

Let's understand the basics:

- Fine-tuning tailors the LLM's weights to specific domains and tasks, improving performance on long-term challenges.

- RAG, on the other hand, injects up-to-date, relevant information at inference time, addressing quick-to-change needs.

The two techniques complement each other brilliantly. For example, a fine-tuned customer support chatbot could leverage RAG to incorporate the latest customer data, providing contextual and personalized responses.

We don't have to pick sides. Embracing the synergy between RAG and fine-tuning can unlock the true potential in real-world applications.

We’re thrilled to introduce AI Agent Templates, a new way to build and deploy AI Agents — directly from your IDE.

Develop agents with your Framework of choice: LangGraph, Llama Index, CrewAI, Bee, AutoGen... and deploy, test, and manage them easily at scale.

We are now introducing new agent templates to simplify even more the process:

1/ Start Locally: Develop and test your AI agent right from your favorite development environment.

2/ Seamless Deployment: Easily deploy your agent to the cloud and access it via an API.

3/ Versatile Applications: Use these templates as a foundation for automation assistants, research bots, and more.

You get a secure and customizable endpoint when your AI service is deployed.

Check the video or get started with this tutorial! https://lnkd.in/gDYDD-cV

This project got over 49,000 stars on GitHub — check Langflow, an open-source, low-code app builder for RAG and multi-agent AI applications.

Here's what you can achieve with Langflow:

🔧 Python-Based and Model-Agnostic: Build workflows that integrate seamlessly with any model, API, data source, or database.

🎛️ Visual IDE: Utilize a drag-and-drop interface to construct and test complex AI workflows effortlessly.

⚡ Playground for Immediate Testing: Instantly test and iterate on workflows with step-by-step control.

🤖 Multi-Agent Orchestration: Manage and coordinate multiple AI agents within your applications.

☁️ Free Cloud Service: Get started in minutes with no setup required.

🚀 Publish and Export: Deploy your applications as APIs or export them as Python applications.

🔍 Observability Integration: Monitor and optimize your workflows with tools like LangSmith, LangFuse, or LangWatch.

🔒 Enterprise-Grade Security: Benefit from robust security and scalability with the free DataStax Langflow cloud service.

🛠️ Customizable Workflows: Tailor workflows or create entirely new flows using Python.

🔗 Ecosystem Integrations: Access reusable components compatible with any model, API, or database.

Langflow is available under the MIT license, and commercial versions are also available.

Give it a try now:
- Get started for free: https://www.langflow.org/
- Documentation: https://docs.langflow.org/
- Repository: https://lnkd.in/gSMce8-b

We 💙 Open Source AI

Happy building, folks!