Tokenizing Text for LLMs, an AI Agent Dictionary, Optimizing Agentic Workflows, and AI for Robotics at ODSC West

"Explore the Future of AI for Robotics at ODSC West Our new AI for Robotics Track explores how AI powers real-world systems from foundation models to autonomous navigation and manipulation. Then put your skills to the test in the Robotics Hackathon, where the winning team takes home a Unitree Go2 Pro Robot Dog. Learn more here A Comprehensive Guide to Tokenizing Text for LLMs Ever wonder how LLMs like GPT turn raw text into something machines can understand?"

"Video of the Week: Differentially-Private Synthetic Data for Everyone with Dr. Michael Platzer In this hands-on session, you'll learn how to generate high-quality synthetic data that preserves privacy using differential privacy techniques. We'll walk through how to train differentially private generative models with MOSTLY AI's open-source Synthetic Data SDK and explore how this method compares to traditional anonymization approaches in terms of both utility and risk."

"Silver Bullets: A Shortcut to Optimized Agentic Workflows The fastest way to stall an agentic AI project is to reuse a workflow that no longer fits. Using syftr, this team identified silver bullet flows for both low-latency and high-accuracy priorities that consistently perform well across multiple datasets. The AI Agent Dictionary: 31 Must-Know Terms Explore the essential terms in our Agentic AI Dictionary covering core concepts, design patterns, frameworks, and safety principles. Stay ahead in 2025 with this must-know guide for AI agents."

ODSC West includes an AI for Robotics track that covers foundation models, autonomous navigation, and manipulation, plus a Robotics Hackathon awarding a Unitree Go2 Pro robot dog. A guide explains tokenization for LLMs and how raw text becomes tokens that preserve context and improve efficiency for NLP tasks. The Silver Bullets piece identifies robust agentic workflows using syftr for both low-latency and high-accuracy priorities across multiple datasets. An AI Agent Dictionary lists 31 essential agentic AI terms, including design patterns, frameworks, and safety principles. MCP defines a metadata protocol for exposing application functionality. A hands-on session demonstrates differentially-private synthetic data training with MOSTLY AI's SDK.

Read at odsc.medium.com