Innovate with Cosmos DB, Semantic Kernel & AI
summary
A working prototype converts short user inputs into art descriptions with GPT‑4o Omni, turns those into vectors using Embeddings 3 Large, and stores them in a single container on Azure Cosmos DB vector preview with vector indexes and partition keys configured; Semantic Kernel orchestrates prompts, embeddings and function-matching. A .NET API and p5.js UI render parametric algorithms, some 60×60 cm prints once took 3–4 hours to plot, similarity search is implemented conceptually and planned but not fully tuned yet.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 6:50Hobby to tech challenge
- 12:31Demo overview and components
- 16:27Cosmos DB deep dive and vector search
- 21:02Semantic Kernel and AI orchestration
- 26:23Demo: Cosmos DB config
- 27:24Backend API workflow
- 31:37AI art descriptions
- 33:48Matching functions to art
- 36:38Frontend UI demo
- 39:27Live demo and interaction
- 42:31Prompt engineering discussion
- 43:15Q&A
related talks
12mBuilding Generative AI Apps with Azure Cosmos DB
25mTurning Data Into Generative Chaos
50mHow to design and build AI applications with vector search using Azure OpenAI & Azure Cosmos DB
46mBuilding AI solutions with Semantic Kernel
01h 16mAzure SQL ❤️ Open AI
26mSemantic Kernel: Teaching AI to Not Just Repeat, But Complete!
48mA Fun & Absurd Introduction to Vector Databases
59mHow to add genuinely useful AI to your webapp (not just chatbots)