Arize AI’s Post

Architects of Tomorrow: Navigating the Generative AI Era! Dive into the future of Software Architecture mastery. Check out my blog post! https://lnkd.in/eupQJVXT

AI can now automate the most repetitive and undifferentiated parts of engineering. One of our most popular automation workflows in early access has been rightsizing Kubernetes pods and nodes through the safe, ubiquitous language of engineers: Github PRs. Here’s a look at how Flightcrew operates like Dependabot to proactively reduce costs and protect uptime on Kubernetes. (read more here! https://lnkd.in/gPnNe44Q)

Something that I am learning right now and enjoying the hell out of is Customized Automated AI System Architecture. Still very much a work in progress but is providing me in the end with a blueprint using the Tools and Resources I already have access to or soon will have access to.

As part of my continuous focus on self-development and learning, I am attending this session to learn more about Software Architecture and Generative AI.

Bringing RAG Architecture to Life with Llama 2, Vector Store, and LangChain 🤖 Our latest guide walks through implementing Retrieval-Augmented Generation (RAG) in practice using Llama 2, Qwak Vector Store, and LangChain. Learn how RAG combines strengths of retrieval and generative models for more relevant, grounded responses. We'll dive into: 🗃️ Indexing documents in Vector Store 💬 Deploying Llama 2 for response generation 🔗 Building a RAG service with LangChain The end result? A contextual chatbot powered by this cutting-edge technique! Check it out here: https://lnkd.in/dDUQnW8w #RAG #LLM #MLOps

The Navie diagram feature visualizes an issue, and Copilot is used to create a tailored solution. Navie AI architect generates code improvements across the code base. Using the Copilot LLM backend helps align with your company compliance and security needs.

Alex Nesbitt joined me to discuss strategy and strategic thinking in this episode of Small Batches. Listen here: https://lnkd.in/gCRB4PJU We discuss different kinds of strategy, the feeling of activities masquerading as strategy, the connection between software architecture and strategy, and, of course, a little of AI. Highly recommend you follow Alex. It's like Small Batches but for strategy!

BREAKING: OpenAI announces Sora, a new text-to-video model. It can generate 1 min long high quality video that already rivals Runway. Sora is currently being tested by experts to identify any risks or problems. No paper yet but here's what we know about the architecture: - Uses a diffusion model, starting with noise and refining it into video. - Builds on transformer architecture for scaling. - Represents videos as patches (similar to tokens in text models), allowing a range of data types. - Incorporates techniques from DALL·E 3 for following text instructions closely.

🤯 Where it all started, must read: The best interactive visualization of Transformer architecture ever made. This tool brings GPT-2 (124M) to life in your browser: • Visualizes embedding, self-attention, and MLP layers • Demonstrates masked attention and positional encoding • Shows Query, Key, Value matrices in action • Lets you adjust temperature for token sampling • Illustrates advanced features like Layer Normalization and Residual Connections This tool offers unprecedented insights into LLM internals. Link in first comment!

Learn how embedding models are built, trained, and used in semantic search systems in this new short course, Embedding Models: From Architecture to Implementation, created with Vectara and taught by Ofer Mendelevitch. Many LLM apps use a single embedding model for both questions and answers. This can lead to issues such as retrieving results that are similar to the question itself rather than relevant answers. With a dual encoder architecture, you can use separate embedding models for questions and answers to better match questions with appropriate answers. In this course, you will use, build, and train a dual encoder model. You’ll also learn: - What are word embeddings, and how they are used - The evolution of embeddings to BERT, where embeddings take into account each word's surrounding context - How a contrastive loss is used to train a dual encoder model with one encoder trained to embed questions and the other responses. - How to analyze a dual encoder’s effect on search relevance and compare it to a retrieval process with a single encoder. Please sign up here! https://lnkd.in/gvYghsxV