As AI continues to evolve, the need to understand and deploy the right models whether large or small has never been more critical. At the recent Microsoft AI Tour, we explored the latest in generative AI, from Large Language Models (LLMs) to Small Language Models (SLMs), and the tools that make them accessible and impactful.
Use Cases: From Automation to Edge AI
Generative AI is transforming industries through:
- Content creation, summarization, and translation
- Customer engagement via chatbots and personalization
- Edge deployment for low-latency, privacy-sensitive applications
- Domain-specific tasks like legal, medical, or technical document processing
LLMs vs. SLMs: Choosing the Right Fit
| Feature | LLMs | SLMs |
|---|---|---|
| Parameters | Billions (e.g., GPT-4) | Millions |
| Performance | High accuracy, nuanced understanding | Fast, efficient for simpler tasks |
| Deployment | Cloud-based, resource-intensive | Ideal for edge and mobile |
| Cost | High compute and energy | Cost-effective |
SLMs are increasingly viable thanks to optimized runtimes and hardware, making them perfect for on-device AI.
Azure AI Foundry: Your AI Launchpad
Azure AI Foundry offers:
- A model catalogue with open-source and proprietary models
- Tools for fine-tuning, evaluation, and deployment
- Integration with GitHub, VS Code, and Azure DevOps
- Support for both cloud and local inferencing
Local AI: The Edge Advantage
With tools like Foundry Local and Windows AI Foundry, developers can:
- Run models on-device with ONNX Runtime
- Use APIs for summarization, translation, and more
- Optimize for CPU, GPU, and NPU
- Ensure privacy, low latency, and offline capability
Customization: RAG vs. Fine-Tuning
| Feature | RAG | Fine-Tuning |
|---|---|---|
| Knowledge Updates | Dynamic | Static |
| Interpretability | High | Low |
| Latency | Higher | Lower |
| Hallucination Risk | Lower | Moderate |
| Use Case | Real-time, external data | Domain-specific tasks |
Both methods enhance model relevance RAG by retrieving external data, and fine-tuning by adapting model weights.
Developer Resources
Get started with:
- Foundry Local SDK
- Windows AI Foundry | Microsoft Developer
- AI Toolkit for VS Code
- Windows ML
- Azure AI Learn Courses
- Join the Azure AI Discord Community
Published Jun 06, 2025
Version 1.0
Lee_Stott
Microsoft
Microsoft