From Zero to One: Building AI Agents From the Ground Up | Federico Bianchi, TogetherAI

Description

🎥 Recorded live at the MLOps World | GenAI Summit 2025 — Austin, TX (October 9, 2025) Session Title: From Zero to One: Building AI Agents From the Ground Up Speaker: Federico Bianchi, Senior ML Scientist, TogetherAI Talk Track: Augmenting Workforces with Agents Abstract: What does it take to design a truly autonomous AI agent—one that can reason, act, and learn independently? In this session, Federico Bianchi, Senior ML Scientist at TogetherAI, walks through how his team builds agents capable of executing end-to-end analytical workflows, transforming raw data into actionable insights. Federico breaks down the core architectural principles behind agent design—covering reasoning, reflection, and structured action—and demonstrates how these ideas come together to form robust, transparent, and adaptable agent systems. He also explores how reinforcement learning, tool integration, and iterative improvement loops enable agents to generalize beyond their initial training and handle increasingly complex tasks. This talk provides both a conceptual roadmap and a technical foundation for those aiming to move beyond demos and build production-grade, self-improving AI agents. What you’ll learn: • Key design principles for building autonomous, production-ready AI agents • How reasoning, reflection, and tool use shape agent performance • The role of reinforcement learning in enabling adaptive, self-improving agents • How to structure workflows that move from raw data to real insights • Why understanding context is critical for effective agent deployment 📍 Recorded: October 9, 2025 — In-Person Session, MLOps World | GenAI Summit 2025, Austin, TX 🔗 Learn more: https://mlopsworld.com #MLOpsWorld #GenAISummit #FedericoBianchi #TogetherAI #AgenticAI #AIinProduction #AIEngineering #ReinforcementLearning #AIAgents #MLOps #LLMOps #AICommunity #GenAI #AugmentingWorkforces #AutonomousAgents