The Top AI Slides from ODSC West 2024

ODSC - Open Data Science
4 min readNov 19, 2024

--

ODSC West 2024 showcased a wide range of talks and workshops from leading data science, AI, and machine learning experts. This blog highlights some of the most impactful AI slides from the world’s best data science instructors, focusing on cutting-edge advancements in AI, data modeling, and deployment strategies. Here’s a breakdown of ten top sessions from this year’s conference that data professionals should consider.

1. Topological Deep Learning Made Easy with TopoX with Dr. Mustafa Hajij

Slides

In these AI slides, Dr. Mustafa Hajij introduced TopoX, a comprehensive Python suite for topological deep learning. Attendees were guided through the capabilities of TopoNetX, TopoEmbedX, and TopoModelX, which extend graph learning to complex topological structures like hypergraphs and cellular complexes. This session demonstrated how to leverage these tools using Python and PyTorch, offering attendees practical techniques to apply in their research and projects. The open-source nature of TopoX positions it as a valuable asset for anyone exploring topological deep learning.

2. Building an Agentic Rag Application with LangGraph with Valentina Alto

Slides

Valentina Alto’s hands-on workshop focused on integrating AI agents with Retrieval Augmented Generation (RAG) to enhance generative AI workflows. The session provided insights into using LangGraph to combine VectorDBs and large language models (LLMs), creating dynamic, adaptive pipelines. Participants learned how multi-agent systems can optimize retrieval processes, providing practical applications for deploying intelligent AI agents across industries.

3. Dimensional Data Modeling in the Modern Era by Dustin Dorsey

Slides

Dustin Dorsey’s AI slides explored the evolution of dimensional data modeling, a staple in data warehousing and business intelligence. Despite the rise of big data technologies and cloud computing, the principles of dimensional modeling remain relevant. This session delved into how these traditional techniques have adapted to data lakes and real-time analytics, emphasizing their enduring importance for building scalable, efficient data systems.

4. Building Reliable Voice Agents with Open Source Tools with Sara Zanzottera

Slides

Sara Zanzottera’s AI slides focused on developing robust voice agents using open-source tools. While LLMs excel in generating text, they face challenges in producing natural, conversational voice interactions. The workshop covered strategies to enhance LLM-powered voice agents, enabling them to handle interruptions, improvise, and maintain conversation flow. This session provided actionable techniques for building responsive voice-based applications.

5. Introduction to Containers for Data Science / Data Engineering with Michael A. Fudge

Slides

Michael A. Fudge’s AI slides introduced participants to using containers in data science and engineering workflows. His session explained how containers can simplify the deployment of data applications by bundling dependencies and configurations into portable images. Attendees gained hands-on experience in creating, configuring, and managing containerized environments, essential for deploying scalable projects in the cloud.

6. A Gentle Introduction to Vector Databases and Their Implementation with Balaji Dhamodharan

Slides

Balaji Dhamodharan’s AI slides offered a deep dive into vector databases, a foundational technology for LLMs. The session walked through practical applications, covering everything from the basics to end-to-end implementations. By the end, participants had a clear understanding of how vector databases can enhance AI models’ performance, particularly in tasks like semantic search and data retrieval.

7. Going From Unstructured Data to Vector Similarity Search with Dr. Steven Pousty

Slides

Dr. Steven Pousty showcased how to transform unstructured data into a vector-based query system. Using real-world examples like images and journal abstracts, the session guided participants through the complete process — from data preparation to deploying vector similarity searches. This hands-on workshop equipped attendees with the skills to optimize data storage for AI applications.

8. Building Reliable Coding Agents with Eno Reyes

Slides

Eno Reyes presented the challenges and opportunities in designing agentic systems for software applications. By drawing parallels with robotics and cybernetics, he illustrated how to develop coding agents capable of making non-deterministic decisions reliably. This session provided insights into creating systems that combine the strengths of multiple AI models for increased robustness and adaptability.

9. Neural Operators: A new era of scientific computing with Dr. Kamyar Azizzadenesheli

Slides

Dr. Kamyar Azizzadenesheli’s AI slides explored the shift from traditional neural networks to neural operators for scientific computing. By focusing on complex domains like weather forecasting, aerodynamics, and material sciences, the session highlighted how neural operators can transform scientific models and simulations. The talk underscored the potential of these advanced techniques to unlock new scientific discoveries.

10. LLMs from Prototype to Production — LLMOps, Prompt Engineering, and Moving LLMs to the Cloud with Sinan Ozdemir

Slides

Sinan Ozdemir’s comprehensive session on LLMOps addressed the practicalities of deploying LLMs like GPT, Llama, and BERT. He delved into strategies for fine-tuning models, optimizing them with techniques like quantization and distillation, and deploying them in the cloud. Attendees walked away with a deeper understanding of the challenges and best practices in moving LLMs from prototype to production.

Conclusion on ODSC West 2024 AI Slides

These sessions from ODSC West 2024 highlighted emerging technologies and practical methodologies that data scientists can leverage to advance their projects. From topological deep learning to deploying LLMs at scale, this year’s conference provided attendees with actionable insights to stay ahead in the fast-paced world of data science and AI.

If you want to go beyond just reviewing slides, then be sure to check out ODSC East May 13th-15th, 2025, in Boston, Massachusetts! It’ll be the 25th anniversary of ODSC East, and we’re planning more networking, more hands-on training, and more surprises than ever before. Register now while tickets are still 75% off!

--

--

ODSC - Open Data Science
ODSC - Open Data Science

Written by ODSC - Open Data Science

Our passion is bringing thousands of the best and brightest data scientists together under one roof for an incredible learning and networking experience.

No responses yet