Implementing LLMs in Enterprise Systems: Practical Strategies and Future Directions

Abstract

This talk provides a roadmap for integrating Large Language Models (LLMs) into business environments. Starting with basic concepts such as LLM APIs and task types, we move towards the strategic application of LLMs in enterprise settings. We discuss the entire process from use case scoping, defining the UI/UX and the LLM system architecture, through to prototyping, and live deployment, highlighting key areas such as dynamic tool selection, testing and evaluation, RAG pre-processing pipelines, and the essential role of context management in defining effective LLM systems. Using examples from actual deployments, this presentation delivers an overview of the key considerations in LLM system development, setting the stage for focused sessions on specific topics later in the day.

Date
Apr 26, 2024 9:15 AM — 10:15 AM
Location
Polytechnique Montreal
2500 Chem. de Polytechnique, Montréal, QC H3T 1J4
Jenn Gamble
Jenn Gamble
Founding Head of Data Science - Distyl AI

Jenn Gamble is part of the founding team at Distyl AI, where she heads Data Science. In partnership with OpenAI Distyl works with Fortune 500s to develop purpose-built AI products that assist and orchestrate operations, preparing enterprises for an AI native operating model. Prior to Distyl, Jenn led the data science team at an IoT engineering firm (Very), and was Director of Data Science at Noodle AI, during the early days of Enterprise AI. Her academic background includes a PhD in Electrical Engineering, and MSc in Statistics.