Fine-Tune SLMs & Build Multi-Agent NL2SQL Applications with Amazon AgentCore

In this hands-on workshop, you'll fine-tune a Small Language Model (SLM) for NL2SQL tasks and deploy it as part of a multi-agent system using Amazon AgentCore.

What you'll build:

• Fine-tune an SLM on a custom NL2SQL dataset using Amazon SageMaker
• Deploy your fine-tuned model as an agent endpoint
• Build a multi-agent architecture with Amazon AgentCore that orchestrates specialized agents (query generation, validation, execution)
• Chain agents to translate natural language questions into accurate SQL queries

What you'll learn:

• SLM fine-tuning best practices (LoRA, dataset preparation, evaluation)
• NL2SQL patterns and prompt engineering for structured outputs
• Multi-agent orchestration with AgentCore (routing, tool-use, memory)
• Cost and latency optimization vs. large foundation models

Who should attend: ML engineers, AI/ML leads, and technical founders at startups who want to build production-grade agentic applications using fine-tuned small models instead of expensive large LLMs.

Practical info:

📍 Station F, Paris

👥 Limited to 15 participants

💻 Bring your laptop — fully hands-on with AWS-provisioned environments

🗣️ Workshop in French, materials in English