Function Calling Updates: What's New in September 2024
Function calling is the foundation of agentic AI. These patterns help you build reliable, type-safe tool integrations that scale.
24 articles
Function calling is the foundation of agentic AI. These patterns help you build reliable, type-safe tool integrations that scale.
JSON schema enforcement transforms unpredictable LLM outputs into reliable, type-safe data that your applications can trust.
Structured Outputs transform unreliable text generation into dependable data extraction. Use them whenever you need guaranteed JSON structure from your AI…
o1 introduces new parameters while removing some familiar ones: Feature GPT 4o o1 preview System messages Yes No Streaming Yes No Function calling Yes No…
API-first data design treats data products as services with well-defined interfaces. Today I'm exploring how to build data APIs in Microsoft Fabric.
Parallel function calling in GPT-4 Turbo is the DevDay improvement to function calling that changes what's architecturally practical with AI agents.…
The seed parameter introduced in GPT-4 Turbo at DevDay 2023 is a useful addition for testing and debugging, but it's important to understand what…
JSON mode in GPT-4 Turbo (gpt-4-1106-preview) is a simple addition that solves a real, persistent annoyance in LLM application development: the model…
Tomorrow we'll explore prompt compression techniques. tiktoken Library OpenAI Tokenizer Azure OpenAI Pricing
TypeScript is a statically typed language that is a popular choice for building large-scale applications. It provides improved reliability, maintainability…
The original API for text generation: Message based API with roles: Create an abstraction that works with both APIs:
I'm very surprised that there isn't a straight forward library out there that is free to use and convenient in terms of file conversion. I think it's…
JSON transcoding bridges this gap.
Rate limiting protects your application by: Preventing denial of service attacks Ensuring fair usage among clients Protecting downstream services Managing…
Long polling is the workaround for near-real-time notifications in environments where WebSocket connections aren't available—the client sends an HTTP…
Feature SSE WebSocket Direction Server to client Bidirectional Protocol HTTP WebSocket Reconnection Automatic Manual Binary data No Yes SSE offers a simple,…
APIM GraphQL support enables modern API patterns with enterprise governance, rate limiting, and authentication.
gRPC offers binary serialization, HTTP/2 multiplexing, and streaming for high performance microservices on Azure.
Custom APIs offer: Better IntelliSense support Strongly typed parameters No workflow dependency Modern development experience Define a custom API for order…
Dataverse offers: Relational data storage Built in security model Business logic capabilities Integration with Power Platform Configure authentication for…
Install and configure the Microsoft Graph SDK: Query and manage user data: Manage calendars and events: Access OneDrive and SharePoint files: Optimize…
A custom connector defines: API endpoints and operations Authentication method Request/response schemas Triggers and actions The easiest approach is…
Minimal APIs in ASP.NET Core 6 are the answer to the question "why do I need controllers, action filters, model binding, and routing tables just to expose a…
Where I actually use gRPC: backend service-to-service in a Kubernetes cluster, where I control both ends and want strong typing across language boundaries.…