Roundtable with Pierre Peigné, Chief Science Officer at PRISM Eval & Alessandro Benedetti, Director at Sease, on Retrieval-Augmented Generation: why it's relevant today, what are the potential benefits and risks, and how to do it properly to capture most of its value in your projects. at DevCon 2024 organised by Algolia
Image Recommendations: the Good, Bad, and Ugly Use-Cases
What data do you use to fuel your product carousels? Textual attributes? Business-relevant metadata? These are all good, but you may have something else valuable: solid product images. But whichs use-cases are a good fit for image-based recommendations? at MICES Conference in Berlin organised by OpenSource Connections
Where's the beef? Evaluating the quality of content in your GenAI Project
So you created a GenAI project based on your company's content and some LLMs. Is the generated content any good? From eyeballing it during development to state-of-the-art user behavior tracking, there are many ways you can approach this. In this talk, you'll se a GenAI feature currently in Private beta at Algolia; how we approached generating content for diverse, global customers; and a few techniques you can try for qualitative and quantitative assessments of your own generative projects! at ParisNLP Meetup
Zucchini or Cucumber? Benchmarking Embeddings for Similar Image Retrieval thanks to your weekly Grocery shopping
In this talk, we present the process of building a benchmark, the results of our evaluation, and the lessons we learned from those to improve our Image Recommendation API. You'll learn what makes a dataset suitable for evaluating a problem, how to pick the right metrics to evaluate, and other useful tips to evaluate your own recommender systems on other specific domains based on any available data! at Haystack US Conference 2024 organised by OpenSource Connections
Visual recommendations: the Good, the Bad, and the Ugly
What makes a recommendation stand out? Are all use-cases a good fit for visual recs? In this talk, Paul-Louis walks us through several real-life examples using LookingSimilar to provide visual recommendations to various kinds of users, teaching what kind of use-cases would benefit most from Image Recommendations, and how to adapt your data to make the most of it! at Algolia DevBit, Spring 2024
ECIR23: Building a Business-Aware Image Retrieval API
In 2023, users expect to be understood across various interaction modes. From YouTube transcript search to Pinterest image match, they are used to search interfaces understanding different modalities beyond pure text. In this talk, I share learnings building an image recommendation API for various business needs, showing how you can leverage powerful computer vision models to serve different goals by packaging it into a flexible API and deploying it to a global audience. at ECIR conference organised by British Computer Society
How to successfully drive a machine learning project ?
Table ronde sur ce qui rend les projets de Machine Learning particuliers : comment la gestion de projet peut s'adapter à la recherche expérimentale, comment parler aux autres acteurs du projet, et comment faire monter en compétence nos collègues. at AI Meetup organised by Ekimetrics
Developer Experience for Machine Learning Engineers
In a Machine Learning product or API, what makes a good Developer Experience? With Anil, we discuss the ingredients of a great ML product; what makes or breaks developer productivity; and how to stay on top of this constantly-evolving field. at Anil Kumar Krishnashetty's Podcast
Craft a Discovery Experience in 30mn with Recommend models
Demo: see how you can create a full content discovery experience, powered by ML Recommendation models, without being a front-end expert. at Algolia DevCon 2022
Entre Devs S01E02: Y'a le bon développeur et le mauvais développeur
Comment parler de notre travail à nos proches ? Les entreprises ont-elle le devoir et la légitimité de changer la société ? Les formations préparent-elles correctement les jeunes développeuses et développeurs ? La maintenance du code, on en parle ? Et la philosophie dans tout ça ? at Podcast `Entre Devs`
How leveraging two APIs (DialogFlow and Algolia) allows to transform a JSON file describing a couple hundreds songs into an interactive, conversational experience. at ParisAPI#31
