AI Tutorials

Follow this Claude Opus 5.5 API tutorial to build a Python incident investigator with vision, programmatic tool calling, counterfactual replay, task budgets, structured outputs, and cost tracking.

Learn to edit images with Qwen Image 2.1 on a Colab A100, guide changes with text or drawings, and export selected results as a GIF.

Follow this Grok 4.7 API tutorial to build a Python circuit reviewer with images, datasheets, code execution, web search, and local verification.

Learn how to stream mic audio, handle interruptions, and call tools asynchronously in Python, then compare Gemini 3.8 Live with its Extended Thinking variant.

Nodes do the work, edges decide what runs next, and shared state carries information between them. This graph engineering tutorial covers when a multi-agent graph actually beats a single loop, and walks through building a researcher, writer, and reviewer pipeline in LangGraph with a conditional retry edge.

Learn how to set up TypeSafe AI's Python SDK, ask Choice, Score, and Noul questions in a single call, build a support ticket router that keeps the routing policy in your own code, and find out where Jev breaks before you ship it.

Build a Python visual repair agent with DeepSeek V4.1 Flash, the Responses API, Playwright screenshots, apply_patch, pytest, context caching, and cost tracking.

Build and run a cloud agent with the OpenAI Agents API that can analyze files, execute code, verify results, and return finished artifacts from a single request.

Run the optimized Motif-3 Quant on an NVIDIA H200 using the ds4 runtime, serve it through an OpenAI-compatible API, and integrate it with a Pi coding agent.

Learn how to set up the Kimi Browser Extension using both Kimi Work and Kimi Code CLI, and run a product search that extracts structured results from a website.

Flow matching trains generative models to turn noise into data by learning a vector field that moves samples along a chosen probability path, the same idea driving continuous normalizing flows and modern diffusion models.

Deep reinforcement learning combines the trial-and-error loop of reinforcement learning with neural networks that generalize across huge state spaces.