Official documentation, by topic
For exploring beyond the course. Required readings and setup steps are in Canvas.
Course references
- Huyen, AI Engineering (O'Reilly, 2025) (opens in a new tab)The course textbook. Free to UNT students through UNT Libraries' A–Z Databases (O'Reilly).
- Model Context Protocol: Tools (opens in a new tab)How a tool standard exposes tools to a model, and what it leaves to you.
- OWASP LLM06:2025 Excessive Agency (opens in a new tab)Excessive functionality, permissions and autonomy.
- OWASP Top 10 for Agentic Applications (opens in a new tab)Goal hijack, tool misuse, privilege abuse and more.
- NIST AI Risk Management Framework (opens in a new tab)The governance framework used in Module 13.
Platforms
- Google AI Studio (opens in a new tab)Browser prototyping of prompts and agent behaviour.
- Gemini API documentation (opens in a new tab)Model capabilities, limits and integration.
- Vertex AI (opens in a new tab)Google's production platform for managed models and deployment.
- Copilot Studio documentation (opens in a new tab)Enterprise agents, connectors and governance.
Models
- Hugging Face documentation (opens in a new tab)Models, datasets and libraries.
- Hugging Face model hub (opens in a new tab)Browse models; check licence, context length, tool calling, cost and safety.
- NVIDIA NIM (opens in a new tab)Production model serving on NVIDIA hardware.
Agent frameworks
- LangChain (opens in a new tab)A common orchestration framework.
- LlamaIndex (opens in a new tab)Retrieval and data frameworks.
- AutoGen (opens in a new tab)Multi-agent workflows from Microsoft.
- CrewAI (opens in a new tab)Role-based multi-agent orchestration.
Retrieval and vector databases
- Pinecone (opens in a new tab)Managed vector database.
- Weaviate (opens in a new tab)Open-source vector database.
- Chroma (opens in a new tab)Lightweight open-source embedding store.
Deployment and serving
- vLLM (opens in a new tab)High-throughput model serving.
- KServe (opens in a new tab)Model serving on Kubernetes.
- Ray Serve (opens in a new tab)Scalable serving and batching.
Evaluation and observability
- LangSmith (opens in a new tab)Tracing and evaluation.
- Arize Phoenix (opens in a new tab)Open-source LLM observability and evaluation.
- TruLens (opens in a new tab)Evaluation framework for LLM applications.
Guardrails and safety
- NVIDIA NeMo Guardrails (opens in a new tab)Programmable guardrails for conversational systems.
- Guardrails AI (opens in a new tab)Validation of model inputs and outputs.