AIF-C01 study notes
- Fundamentals of AI and ML AIF-C01 notes on AI, ML and deep learning terms, data types, learning styles, inferencing modes, when ML fits, the AWS managed AI services, the ML lifecycle on SageMaker, MLOps and evaluation metrics. 20%
- Fundamentals of generative AI AIF-C01 notes on tokens, embeddings, transformers, diffusion, foundation models and their lifecycle, generative AI use cases, strengths and limits, model selection factors, business metrics, and the AWS services and cost model for generative AI. 24%
- Applications of foundation models AIF-C01 notes on selecting a pre-trained model, inference parameters, RAG and vector stores, Bedrock Knowledge Bases and Agents, customisation cost tradeoffs, prompt engineering techniques and risks, fine-tuning data preparation, and evaluation metrics such as ROUGE, BLEU and BERTScore. 28%
- Guidelines for responsible AI AIF-C01 notes on the dimensions of responsible AI, Bedrock Guardrails, legal risks, dataset characteristics, bias and variance, bias detection tools (Clarify, Model Monitor, A2I), transparency and explainability, model cards and human-centred design. 14%
- Security, compliance and governance for AI solutions AIF-C01 notes on securing AI systems on AWS (IAM, KMS, PrivateLink, Macie), data provenance and lineage, secure data engineering, AI-specific threats such as prompt injection, compliance standards, governance services and frameworks like the Generative AI Security Scoping Matrix. 14%