MLOps Community San Francisco Fall & Winter Workshops - Part IV
In Part IV of our "Fall & Winter Workshops" series, the MLOps.community is partnering with Qdrant (a high-performance vector database) to organize a half-day hands-on workshop on Feb 21st, 2024 where you'll learn how to optimize your RAG app for production. These workshops are for Software Engineers, AI/ML Engineers, and Product Managers trying to figure out how they productize the GenAI prototypes they've built.
This will be a half-day event with two sessions:
1:00pm - 3:00pm: Keeping up with OpenAI et al. by Rahul Parundekar at the MLOps Community.
3:00pm - 3:30pm: Break
3:30pm - 5:30pm: Advanced RAG strategies: optimizing your semantic retrieval by Qdrant.
Session I: Keeping up with OpenAI et al. by Rahul Parundekar at the MLOps Community.
Objectives: In the fast-paced world of RAGs, new LLMs, prompt strategies, and embedding models are released almost every week. Just this past month, OpenAI launched its new embedding models. Is it time for you to switch? What are the implications of switching to newer models for your user? What does Ops look like to deploy these changes to production?
In this workshop, you’ll gain the skills to iteratively improve your RAG performance with a solid experimentation methodology. We'll start by choosing a "north star" metric for your RAG system performance. We'll then learn how to track and make improvements to the prompts, embedding models, and LLMs. Finally, we'll cover the MLOps challenges and strategies to roll out changes to production.
🔥 📣 I'm thrilled to announce that Shahul ES and Jithin James, the creators of the Ragas score (docs.ragas.io), will be explaining it in person at our workshop next week! Shahul and Jithin, who are currently in SF since they are part of the current YC batch, will be speaking during the workshop on "Keeping up with OpenAI et. al" (Session I). They'll cover the metrics and also share some real-life conversations that motivated them to build this.
Target Audience This course is designed for Software Engineers, ML Engineers, and Data Engineers who want to learn the latest techniques for improving your RAG system.
Prerequisites:
Proficient in Python
Basic understanding of RAG.
Session II: Advanced RAG strategies: optimizing your semantic retrieval by Qdrant.
Description: LLM outputs are only as good as the documents we provide for answer generation. We’ll start with an existing RAG pipeline using Qdrant and go through advanced strategies for optimizing it. Optimization constraints depend on your speed, memory, and quality requirements - pushing semantic search to the limits and combining it with different retrieval methods.
Instructor: Kacper Lukawski is a software developer and data scientist at heart, with an inclination to teach others. Public speaker, working in DevRel.
Format: Hands-on and code walkthrough
Who it is for Developers planning to build or already building Retrieval Augmented Generation applications or anyone who wants to optimize retrieval in GenAI.
What you will learn:
Basics of vector search.
The challenges of vector search based on neural embeddings.
Tweaking semantic retrieval.
What to do when you need to increase search quality, reduce memory requirements, or improve speed.
Building hybrid search. Mixing different retrieval strategies to handle scenarios in which vector search fails.
Prerequisites:
Some experience in implementing RAG, with or without high-level frameworks such as Langchain or LlamaIndex
Basic familiarity with vector search and information retrieval concepts
Resources: Slides and Jupyter notebooks
**PLEASE BRING YOUR OWN LAPTOPS**
Thanks to the amazing folks at Microsoft Reactor for being a community partner and hosting our upcoming events!!
Note: To comply with the venue, we've added a few questions that have been requested of us. Thank you for understanding.