AI Engineering Specialization
Curriculum Readiness: Learner-ready
The complete teaching, practice, assessment, remediation, and evidence path is available. Editorial, technical, accessibility, security, external, and learner-pilot review remain tracked separately from curriculum readiness.
Goal
Develop an AI-enabled system whose value, failure behavior, safety, latency, and cost can be defended with reproducible evidence. An API-only demonstration without a baseline, evaluation set, failure analysis, and fallback does not pass.
Prerequisites
- Semester 1 probability, statistics, and linear algebra
- Semester 2 algorithms
- Semester 6 data and distributed systems
- Semester 9 delivery, security, and observability
Modules
- Machine-Learning Foundations and Baselines
- Neural Networks, Embeddings, and Transformers
- Retrieval, Ranking, and Data Pipelines
- Evaluation, Safety, and Security
- Serving, Observability, Cost, and Operations
Graduation Evidence
- dataset and provenance record
- non-AI baseline
- reproducible training or evaluation pipeline
- evaluation set and failure taxonomy
- model/system card
- adversarial and prompt-injection tests
- latency and cost analysis
- human escalation and fallback design
- NIST AI RMF risk register
- public-safe case study and oral defense
Use the project brief, checkpoint, and rubric as the binding gate.