Gustavo Penha

I am a research scientist at Spotify. I have a PhD from TU Delft in conversational search and recommendation. My research interests include information retrieval, recommender systems, natural language processing, and machine learning. I am currently focused on recommendation and search problems. In this context, some techniques and topics that I’ve worked with are dense/generative recommendation & retrieval models, re-rankers, LLM-as-judges, alignment, explainability, negative sampling, synthetic data generation, knowledge distillation, user profile generation, and RAG.

news

May 26, 2023 Defended my PhD on Conversational Search and Recommendation 🎓.
Jan 2, 2023 Joined Spotify as a research scientist 💻 🎧.
Apr 19, 2022 We won the best paper award with our ECIR’22 paper 🏆 🎉.
Aug 27, 2021 Finished my research internship at Amazon 📦: CHI’22 & CHIIR’22.

selected publications

  1. RecSys
    Bridging Search and Recommendation in Generative Retrieval: Does One Task Help the Other?
    Penha, Gustavo, Vardasbi, Ali, Palumbo, Enrico, Nadai, Marco, and Bouchard, Hugues
    PDF Blog Slides
  2. ECIR 🏆 best paper
    Evaluating the Robustness of Retrieval Pipelines with Query Variation Generators
    Penha, Gustavo, Câmara, Arthur, and Hauff, Claudia
    PDF Code Slides
  3. RecSys
    Probing BERT for Conversational Recommendation
    Penha, Gustavo, and Hauff, Claudia
    PDF Code Slides

presentations

Slides for our RecSys'24 paper on generative search and recommendation.
Slides for our WebConf'23 paper on retrievability of entities.

academic services

  • Reviewer for SIGIR (21, 22, 23, 24), RecSys (21, 22, 23, 24), ECIR (19, 20, 21, 22, 23, 24), CIKM (19, 21, 22, 23), TheWebConf (20, 23).
  • Organizer at Search-Oriented Conversatinal AI workshop (SCAI) at SIGIR'22.

one page cv