Relari

"Is RAG all you need?" - check out this talk by Sara Zanzottera at EuroPython, where she highlighted our integration with Haystack, the popular LLM framework by deepset.   Some highlights: 1️⃣ How RAG works and how it fails 2️⃣ ChatGPT (without RAG) vs Perplexity (with RAG) 3️⃣ How to use continuous-eval by Relari to evaluate RAG performance objectively   Check out the presentation, slides, and example code in this link: https://lnkd.in/erqypcAr

Are you getting the most out of your LLMs and RAGs? In the latest episode of the ODSC Ai X Podcast, join Pasquale Antonante, PhD, Co-founder and CTO at Relari AI, to discuss innovative evaluation methods for LLM and RAG applications. Get ready to explore: ✅ Measuring LLM Performance: The different ways to evaluate LLM pipelines  ✅ Understanding RAGs: Key evaluation metrics and how to measure precision, recall, faithfulness, relevance, and correctness ✅ The Human in the Loop: The role human evaluation plays in assessing the quality of generated text ✅ The Role of Synthetic Data: Using synthetic data based evaluation for LLMs and RAG and leveraging synthetic data for LLM fine-tuning 🎧 Listen Now:  https://linktr.ee/odsc #AI #LargeLanguageModels #RAG  #Podcast #ODSC

I was inspired by the SF community breakfasts hosted by Inferless and started organizing "GenAI Developer Breakfast" at Harvard Square while we are in Boston. I find these intimate small-group conversations super fun and insightful. Thank you, everyone, for coming today and exchanging learnings on building LLM-powered applications across a variety of industries! Jon Duffy, Babak Kia, Chris Tierney, Axel Solano, umamaheswar E., Tyce Herrman, Gabor Soter

Earlier this month, I had a great conversation with Sheamus from Open Data Science Conference (ODSC). We discussed RAG evaluation, the challenges of evaluating complex pipelines and how to effectively leverage synthetic-data.  I also shared some of our latest learnings from working with top AI teams deploying LLMs in production. Our chat is out now in the latest Ai X podcast episode! Check it out here: 🍏 Apple Podcast (https://hubs.li/Q02Cvbdx0) 🎧 SoundCloud (https://hubs.li/Q02CvbXF0) 🎵 Spotify (https://hubs.li/Q02Cv9Wb0) 📦 Castbox (https://hubs.li/Q02CvcmW0) Reach out to me if you have any questions or want to learn more about our work at Relari. #AI #MachineLearning #DataScience #ODSC #AiX #RAG #Evaluation #LLM #GenAI #SyntheticData

Terrific night last night! Thank you to our speakers Ashish Bhatia, Suman Debnath, and Yi Zhang. Three independently great talks that complemented each other so well. Thanks to my cohosts Dan Elton, Ph.D. and Benjamin Chen 👏 and of course Bill Liu who is behind it all. Interested in speaking or getting involved? DM me! https://lnkd.in/ewrP5XCn

I'm very pleased to announce the speaker lineup for the June AICamp Boston meetup at Microsoft New England Research & Development Center (NERD)! The meetup will be held next Tuesday, June 18th. RSVP here to attend: https://lnkd.in/eAfxZdMq -------- "Tailoring Foundation Models for Generative AI Applications" ----- Speaker: Suman Debnath (Amazon Web Services (AWS)) Abstract: In this session, we will focus on fine-tuning, continuous pretraining, and retrieval-augmented generation (RAG) to customize foundation models using Amazon Bedrock. Attendees will explore and compare strategies such as prompt engineering, which reformulates tasks into natural language prompts, and fine-tuning, which involves updating the model's parameters based on new tasks and use cases. The session will also highlight the trade-offs between usability and resource requirements for each approach. Participants will gain insights into leveraging the full potential of large models and learn about future advancements aimed at enhancing their adaptability. --------------- "Building AI Applications Responsibly" --------------------- Speaker: Ashish Bhatia (Microsoft) Abstract: In this session, I will walk through the audience on what it takes to AI application responsibly within Enterprises.. --------------- "Generating Synthetic Data to Test LLM Applications" ------- Speaker: Yi Zhang (Relari) Abstract: In this session, I will discuss the benefits of using synthetic datasets, how to generate high-fidelity synthetic data and examples for RAG and AI agent applications. Once again, thank you Bill Liu for your help. You can find a comprehensive listing of in-person Boston-area AI events here: https://lnkd.in/eCNfzpWr

I'll give a talk on synthetic data at the Unstructured Data Meetup at GitHub's SF HQ tomorrow night. You'll also hear from Jiang Chen, Head of Ecosystems & AI at Milvus, Michael Del Balso, CEO of Tecton, and Chaoyu Yang, CEO of BentoML across many LLM / RAG topics. Come join us if you are in town, or watch the live-stream on twitch!

Check out this case study on using synthetic data to test RAG systems with Relari, LangChain and #Milvus Lite. Dig in 👇

🚀 Check out our latest case study where Yi Zhang and Pasquale Antonante walked through an end-to-end example to: 0️⃣ Build 3 Simple RAG systems (or bring your own RAG) 1️⃣ Generate Synthetic Dataset 2️⃣ Systematically benchmark the RAG systems with the synthetic dataset 🙌 We're excited to collaborate with Milvus on this case study! Our synthetic data pipeline has been integrated with the new lightweight Milvus Lite vector database released today! Check out the post for details or try for yourself in the notebook. #syntheticdata #generativeai #enterpriseai #llm #llmevaluation

Relari provides data-driven development infrastructure for GenAI applications, and they're hiring a founding engineer. I was impressed by how quickly Yi Zhang was able to execute during YC and convert impressive logos from pilots to MRR.