PyData Montreal Meetup #26 (in-person | en personne)

New year, new PyData meetup! 🎉

📢 The MontrealDataCommunity is proud to present out first event of 2024 (with many more to come) at Potloc.

We’re starting the year big with an incredible roster of speakers:

• Zakaria Sassioui, from LaPresse
• Jean-Philippe Corbeil & Hadi Abdi Ghavidel, from Microsoft & Bell respectively

As well as a live demonstration from 2 of our very own MontrealDataCommunity organizers from Potloc,

• Jean Philippe Petit-Frère & Vinh Nguyen

We are very excited to be having this first meetup of the year at Potloc HQ (300-5420 St Laurent Blvd, Montreal, Quebec H2T 1S1).
IMPORTANT :

• The access point is on 5420 St Laurent Blvd, office #300, Montreal, Quebec H2T 1S1
• Get your proof of attendance with you (available on the meetup event page or received per email), for us to check your status (waiting list or not) at the venue.

📊🤖 Don’t miss this opportunity to learn from some of the brightest minds in the Montreal data space, as well as meet fellow data practitioners & enthusiasts. We’re excited to see you there l! 👋😄

AGENDA

• 17h30 - Open doors
• 17h50 - Introduction
• 18h00 - Talk #1 - Understanding semantic search
• 18h30 - Break
• 18h45 - Talk #2 - An Exploration with Extractive Question-Answering Transformers by Bell Canada NLP-Speech Team
• 19h15 - Break
• 19h30 - Talk #3 - Introduction to llm fine-tuning and deployment
• 20h00 - Networking
• 21h00 - End of event

TALKS

1. Understanding semantic search
By Zakaria Sassioui

Description of the talk :
Semantic search is a technology that changes the way we interact with information. This presentation offers an overview of semantic search and how to develop it.
Participants will leave with a solid understanding of how semantic search works and differs from traditional search methods,
with practical advice for implementing powerful semantic search tools.

2. An Exploration with Extractive Question-Answering Transformers by Bell Canada NLP-Speech Team
By Jean-Philippe Corbeil & Hadi Abdi Ghavidel

Description of the talk:
This talk introduces a pipeline tailored for NLU (natural language understanding) use cases, particularly in customer service like at Bell Canada. The pipeline extracts a taxonomy of intents from noisy conversational data using a Question-Answering model and advanced clustering techniques. Our approach achieves 94.3% accuracy in reconstructing intent schemes across diverse domains. Attendees will gain insights into how this semi-automated technique enhances the design of intent taxonomies, particularly when dealing with the intricacies of real-world dialogues.

3. Introduction to llm fine-tuning and deployment
By Jean Philippe Petit-Frère & Vinh Nguyen

Description of the talk:
Get ready for a hands-on journey into the world of Large Language Models (LLMs)! In this meetup, we’ll break down the essentials of fine-tuning, making it easy to grasp for all levels of expertise. From understanding the basics to nailing the fine-tuning process, we’ve got you covered. But that’s not all – we’ll show you how to take your finely-tuned models and deploy them effortlessly on Amazon SageMaker. No more mysteries around the deployment phase – we’re here to guide you through each step.

We’re excited to see you there 👋