This hands-on workshop offers a practical deep dive into Sakura, a modular, on-premise platform for deploying and orchestrating Generative AI models in full privacy.
What You’ll Learn
The workshop is structured to give participants both conceptual understanding and real-world skills in running private GenAI solutions. Attendees will:
- Understand the Architecture of Sakura: Learn how Sakura integrates LLMs, a model hub, retrieval-augmented generation (RAG), multimodal capabilities, and agents in a cohesive framework.
- Integrate and Orchestrate Language Models: Learn how to discover, pull, and run various LLMs (e.g., LLaMA, Phi-3, Mistral) within Sakura using the model hub.
- Create and Interact with Personas: Use Sakura’s framework to build digital personas with persistent memory, embeddings, and prompt chains. Practice injecting memory traces and customizing behavioral parameters for simulation or UX testing.
- Explore Use Cases: Engage with practical examples of Sakura’s power in:
- Code generation and documentation assistance
- Natural language data queries
- Visual question answering and image captioning
- Voice-to-text conversations with AI agents
- Lightweight deployment on edge devices
What You’ll Do
Through guided labs, participants will:
- Explore how to run local LLMs securely without cloud dependencies
- Connect via Web to models via Sakura’s orchestration layer
- Implement a simple chatbot
- Simulate a customer persona and analyze interaction data