Posts Tagged ‘DevoxxUkraine’
[DevoxxUA2023] Orchestrate Your AI with Semantic Kernel
Soham Dasgupta, a Cloud Solution Architect at Microsoft, presented an illuminating session at Devoxx Ukraine 2023, titled Orchestrate Your AI with Semantic Kernel. With over 16 years of experience in software development, Soham demystifies the complexities of integrating AI into applications using Microsoft’s Semantic Kernel SDK. His talk, featuring live coding, provides developers with practical tools to harness large language models (LLMs), aligning with the conference’s focus on AI-driven development.
Understanding Semantic Kernel
Soham introduces Semantic Kernel, an open-source SDK designed to simplify the integration of LLMs into applications. He explains that Semantic Kernel acts as an orchestrator, enabling developers to manage AI requests and responses efficiently. Unlike other frameworks like LangChain, which supports a broader range of LLMs, Semantic Kernel is optimized for Azure Open AI and select models, offering a streamlined approach for Java developers.
Through a live coding demo, Soham demonstrates how Semantic Kernel reduces code verbosity, allowing developers to focus on business logic and prompt design. He showcases a simple application connecting to an LLM, highlighting how the SDK abstracts complex interactions, making AI integration accessible even for those new to the technology.
Simplifying AI Integration
Delving deeper, Soham illustrates how Semantic Kernel enables modular application design. Developers can create objects to connect to specific LLMs, ensuring flexibility without overwhelming complexity. He advises against combining multiple LLMs in a single application, as their non-deterministic nature can introduce unpredictability. Instead, Soham recommends selecting a framework that aligns with the chosen LLM, such as Semantic Kernel for Azure Open AI or LangChain for broader compatibility, including Hugging Face or LLaMA.
His demo emphasizes practical use cases, such as generating context-aware responses for conversational applications. By leveraging Semantic Kernel, developers can orchestrate AI workflows efficiently, reducing development time and enhancing application responsiveness, a key theme of the conference.
Addressing Data Privacy Concerns
Soham addresses a critical concern raised throughout the conference: data privacy in AI applications. He explains that Azure Open AI ensures data remains within a user’s subscription, offering robust privacy controls for enterprise use. In contrast, public versions of LLMs, like Open AI’s standard endpoints, may use data for further training unless an enterprise version is selected. Soham urges developers to read the fine print before integrating LLMs, as sensitive data, such as customer contracts, could inadvertently contribute to model training if not handled properly.
This focus on privacy aligns with the conference’s discussions on ethical AI, providing developers with practical guidance to navigate enterprise requirements while leveraging AI’s capabilities.
Practical Recommendations for Developers
Soham concludes by encouraging developers to explore Semantic Kernel through Microsoft’s Azure platform, which offers resources for hands-on learning. His live coding demo, despite the pressure of a live audience, showcases the SDK’s ease of use, inspiring developers to experiment with AI orchestration. Soham invites further engagement via social platforms, sharing his enthusiasm for building AI-driven applications that are both powerful and responsible.
His presentation, delivered with clarity and technical expertise, equips developers with the tools to integrate AI seamlessly, fostering innovation while addressing practical and ethical considerations.
Links:
Hashtags: #AI #SemanticKernel #LargeLanguageModels #Azure #Java #Microsoft #SohamDasgupta #DevoxxUkraine2023