How amd-enterprise-ai/solution-blueprints works
Project Type & Purpose
This repository is a collection of AMD “Inference Microservice Solution Blueprints” (AIMSB). It isn’t one monolithic app but a set of self-contained microservice blueprints (e.g., agentic-testing, llm-chat, talk-to-your-documents) and the underlying Helm charts (under aimcharts) that they rely on (vector DB, embedding service, LLM service). Each blueprint packages its own source code, Dockerfiles (where needed), and Kubernetes Helm templates for deployment.
Key Architectural Layers & File-Structure Indicators
• Top-level:
– aimcharts/… → Helm charts for shared services:
• aimchart-chromadb (vector database)
• aimchart-embedding (embedding model service)
• aimchart-llm (large language model service)
– solution-blueprints/… → individual solution blueprints. Each subfolder:
• Chart.yaml, values.yaml & templates/… → Helm chart for that microservice
• src/ or app/ → the microservice’s code (Python agents, APIs, UIs)
• docs/ → deployment instructions
• pyproject.toml, README.md, LICENSE → project metadata, licensing
Additional Insights from README
This is a versioned collection of example microservice “patterns” that AMD provides for inference workloads. There is no single flow; instead each blueprint demonstrates a particular use case integrating user-code + AIM charts.