Complementary items recommendations using a RAG approach | DSPT & Porto Tech Hub
Wednesday 29 January 2025
Starts 18:30 PM
Finishes 20:30 PM
Organized by Data Science Portugal (DSPT)
Venue: Porto Innovation Hub
Address: Rua Dr. António Luís Gomes
meetup2 Porto
About this event
Hello DSPT community!
Porto Tech Hub is celebrating 10 years! Join us on January 29th at Porto Innovation Hub, and lets celebrate together!
We’ll have a talk by [Tiago Cabo](https://www.linkedin.com/in/tiagocabo/), Data Scientist @ OLX group and [José Medeiros](https://www.linkedin.com/in/joseprsm/), Machine Learning Engineer @ OLX group about an use case of RAG for recommendations.
Don’t miss this opportunity to get to meet the community and deepen your knowledge on the topic.
Reserve your spot today, and don’t forget to add the event to your calendar!
=== SCHEDULE ===
The preliminary agenda for the meetup is the following:
18:30 — 18:45: Get together.
18:45 — 18:55: Welcome message.
19:00 — 19:40: Talk + Q&A: **“Complementary items recommendations using a RAG approach- an OLX use case”** by **[Tiago Cabo](https://www.linkedin.com/in/tiagocabo/)**, DS @ OLX and **[José Medeiros](https://www.linkedin.com/in/joseprsm/)**, MLE @ OLX group
19:40 — 20:30: Networking.
=== Talk Abstract ===
* **Complementary items recommendations using a RAG approach- an Olx use case**
In this presentation, we will explore a cutting-edge approach to generating complementary recommendations for e-commerce items by leveraging large language models (LLMs) and semantic search. We will outline how LLMs are used to generate candidate recommendations and how semantic search enhances retrieval efficiency. A critical part of our process involved creating an offline dataset to evaluate and select the best embedding models for these tasks, incorporating domain-specific factors such as product categories and user preferences. Additionally, we introduce a reranking mechanism based on location proximity to optimize recommendations for localized contexts. Finally, we will discuss the challenges and strategies for productizing this pipeline, ensuring scalability and maintaining high performance in real-world scenarios.
=== About the speakers ===
* **Tiago Cabo DS@OLX group**
I graduated in Mechanical Engineering in 2017 and later transitioned into data science. My first significant professional experience was in novelty and anomaly detection, where I specialized in unsupervised learning techniques. I then shifted my focus to recommendation systems, working at Farfetch, where I developed personalized solutions for users. After two years, I joined OLX, where I continue to explore innovative approaches to recommendations, including the integration of LLM agents and advanced recommender systems. Outside of work, I enjoy outdoor activities, reading, and spending quality time with my baby girl.
* **José Medeiros MLE@OLX group**
I’m José, a Machine Learning Engineer specializing in MLOps. I’m a terrible data scientist and just an OK programmer, but I’ve gotten pretty good at making ML systems actually work… sometimes.I’m also doing a PhD in Psychology, focusing on consumer behavior, while also dabbling in painting to add some color to life.
When I’m not wrangling pipelines or pondering why people buy things they don’t need, you’ll find me outdoors with my kid, proving that work-life balance is achievable— at least until naptime ends.
This page last updated Monday 30 December 2024 at 02:45.
Problems? Report an error or inappropriate listing here.
Information displayed here is provided in good faith but we are not responsible for the content of any listing. Sometimes events can be cancelled or changed at short notice. Please check with the venue or organizer before you travel!