Language Models for Code Completion: A Practical Evaluation

Join one of our regular open reading club sessions to discuss a scientific paper from the ICSE '24 conference on "Language Models for Code Completion: A Practical Evaluation" alongside one of its authors, Jonathan Katzy, PhD candidate from the Software Engineering Research Group (SERG) at TU Delft.

We'll discuss a (rare) academic study that includes both offline and online evaluations for a ubiquitous feature of modern Developer Tools - code completion, when that is backed by pre-trained large language model (LLM). It compares several OSS LLMs in a single-line code completion settings and uses actual IDE plugins to collect real auto-completion usage data from more than 1200 users for over a year.

Join us for a short presentation of the paper and an informal Q&A with the author afterwards! 🧑‍🔬🍕🍺

This event will be held in a hybrid format at the JetBrains Amsterdam Terrace Tower office and will also be streamed online.