Build your understanding.
Advance your practice.
Explore a progression from AI foundations to applied integration and advanced systems work.
These 20 syllabi extend the existing course calendar. Open a course to see its prerequisites, learning outcomes, planned activities and capstone. Learner study editions are available for NTL-111, NTL-212 and NTL-312, including lessons, practice quizzes, workbooks and capstones. Other teaching packages remain in development. Enrollment and start dates have not been announced.
20 of 20 courses shown
No courses match these filters. Reset filters to view the full calendar.
Year 1 · Foundations
Build AI literacy, evidence checking, data reasoning and safe everyday workflows.
NTL-111 AI Literacy & Intelligent Systems Foundations
Understand core AI concepts, capabilities, limitations and responsible use.
Learner study edition available · Enrollment closed
Open the complete learner study edition — lessons, practice quizzes, workbook, capstone and final assessment tasks. Instructor-reviewed enrollment remains closed.
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: No prior AI coursework required. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Distinguish prediction, generation, retrieval and rule-based automation.
- Select a bounded AI task and define evidence of success.
- Check a generated answer against supplied facts and record uncertainty.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. AI and intelligent systems | Classify twelve everyday tools by their function and explain two ambiguous cases. | 75 |
| 2. How generative systems use context | Compare two answers to the same task with and without a source note. | 75 |
| 3. Task selection and human judgment | Score five workplace tasks for suitability and identify a human decision owner. | 75 |
| 4. Evidence and hallucination checks | Build a claim-to-source table and correct three unsupported statements. | 75 |
| 5. Responsible information handling | Redact a fictional customer record and define permitted use. | 75 |
| 6. A repeatable learning workflow | Run a small assistant task, inspect the output and document the corrections. | 75 |
| Capstone | A verified briefing workflow with source notes, five evaluated outputs and a human review checklist. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
No coding required. Use a text editor and an AI assistant already available to you.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-112 Prompt Engineering Foundations
Write structured prompts, evaluate outputs and build reusable prompt patterns.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-111. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Write a prompt with explicit input, output and missing-data rules.
- Distinguish conflicting requirements from competing preferences.
- Compare prompt versions against a fixed test set.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Task and audience | Turn an ambiguous request into a testable task specification. | 75 |
| 2. Facts and context | Separate instructions from source material and define missing-information behavior. | 75 |
| 3. Examples and output contracts | Write three example pairs, including an incomplete input. | 75 |
| 4. Conflicts and priorities | Resolve four conflicting requirement pairs using explicit priority or clarification. | 75 |
| 5. Testing and revision | Compare a baseline prompt with three documented revisions. | 75 |
| 6. Prompt ownership | Create a versioned library entry with limitations and a review date. | 75 |
| Capstone | A source-grounded prompt with ten test cases, expected and actual outputs, revision notes and a library entry. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
No programming prerequisite. The recovered AA-PEP-1CR-001 professional workbook is an extension, not an automatic duplicate-credit award.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-113 Generative AI for Research & Knowledge Work
Use AI for research planning, synthesis, source checking and knowledge workflows.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-111, NTL-112. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Break a research question into answerable subquestions.
- Trace material claims to primary sources.
- Separate evidence, inference and uncertainty in a short brief.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Research scope | Define a decision, audience and exclusions before searching. | 75 |
| 2. Search planning | Draft a search log with alternative terms and stopping criteria. | 75 |
| 3. Source evaluation | Compare authority, recency, methods and limitations of four sources. | 75 |
| 4. Extraction and provenance | Capture claims with titles, dates, links and relevant passages. | 75 |
| 5. Synthesis without fabrication | Reconcile two conflicting sources without hiding the disagreement. | 75 |
| 6. Briefing and verification | Audit a final brief claim by claim and remove unsupported statements. | 75 |
| Capstone | An evidence-backed research brief with a source log, claim matrix and unresolved-question register. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
A browser and word processor are sufficient. Access dates belong in the evidence log.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-114 Data Literacy for AI
Read, clean, interpret and question data used in AI-assisted decisions.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-111. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Identify row meaning, field types and missing values.
- Calculate simple summaries with explicit denominators.
- Detect leakage, biased samples and unsupported causal claims.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Data questions and grain | Define what one row represents in a fictional sales table. | 75 |
| 2. Types and missingness | Create a data dictionary and distinguish zero from unknown. | 75 |
| 3. Cleaning and reconciliation | Flag duplicates and reconcile totals before and after cleaning. | 75 |
| 4. Summaries and visual choices | Compute counts, rates and medians and select an appropriate chart. | 75 |
| 5. Sampling and leakage | Explain how a convenience sample or future information can distort conclusions. | 75 |
| 6. Communicating limitations | Write a short findings memo with caveats and reproducible calculations. | 75 |
| Capstone | A cleaned small dataset, data dictionary, reproducible analysis and a limitations memo. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Use a spreadsheet or Python. Work only with supplied fictional or authorized non-sensitive data.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-115 Python Foundations for AI
Build basic Python literacy for data, automation and AI workflows.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-114. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Use variables, lists, dictionaries and functions in a small script.
- Read CSV or JSON and validate required fields.
- Handle predictable failures and record reproducible output.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Values and control flow | Transform a list using explicit conditions and loops. | 75 |
| 2. Functions and contracts | Write a function with documented inputs, outputs and failure behavior. | 75 |
| 3. Files and structured data | Read a fictional CSV and serialize a validated JSON record. | 75 |
| 4. Cleaning and calculations | Normalize missing values and compute a checked summary. | 75 |
| 5. Errors and tests | Test empty input, malformed rows and boundary values. | 75 |
| 6. Reproducible scripts | Run a small pipeline from a documented command and save its output. | 75 |
| Capstone | A runnable local data-cleaning script with sample input, expected output and five boundary tests. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Use a supported Python 3 installation. Exact environment versions must be recorded with submitted work.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-116 Responsible AI & Digital Safety
Recognize privacy, bias, security and responsible-use requirements.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-111. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Minimize data sent to an AI service.
- Recognize untrusted instructions and permission boundaries.
- Document a risk, its control, owner and residual limitation.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Data minimization | Remove unnecessary identifiers from a fictional intake record. | 75 |
| 2. Bias and affected users | Compare failure consequences for two user groups. | 75 |
| 3. Trusted instructions and untrusted content | Identify an injected instruction inside a document and keep it as data. | 75 |
| 4. Human review and escalation | Define who approves an external message or irreversible action. | 75 |
| 5. Evidence and incident response | Write an incident record using observed facts and unknowns. | 75 |
| 6. Responsible release | Assess a bounded prototype against a risk register and release checklist. | 75 |
| Capstone | A responsible-use plan for one AI workflow, with redaction examples, risk controls and an escalation path. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
This course teaches operational controls. Any jurisdiction-specific legal claim requires separately dated primary-source verification.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-117 AI Productivity & Workflow Fundamentals
Design reliable everyday AI workflows with human verification.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-112, NTL-116. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Map a recurring task from trigger to reviewed output.
- Measure baseline effort and review cost.
- Document a workflow another person can repeat.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Workflow boundaries | Map a task with inputs, owner and completion evidence. | 75 |
| 2. Prompted work stages | Separate extraction, drafting and review. | 75 |
| 3. Templates and handoffs | Create an input form and a structured handoff record. | 75 |
| 4. Quality and exception handling | Route missing data and low-quality outputs to human review. | 75 |
| 5. Time and quality measurement | Compare observed task time and error rates against a baseline. | 75 |
| 6. Operating instructions | Run the workflow twice and revise its operating procedure. | 75 |
| Capstone | A tested everyday workflow, an operating procedure and a measured before-and-after comparison. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Local drafts are sufficient. No messages, purchases or production changes are required.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
Year 2 · Applied integration
Connect tools, data and services. Build a working system and test how it fails.
NTL-211 Retrieval-Augmented Generation Systems
Design RAG pipelines using retrieval, embeddings, grounding and evaluation.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-113, NTL-115. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Build a small retrieval pipeline with source identifiers.
- Separate retrieval quality from answer quality.
- Handle questions the knowledge collection cannot answer.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Problem and corpus | Select a bounded document set and write ten answerable and unanswerable questions. | 75 |
| 2. Chunks and metadata | Create chunks with stable document, section and date identifiers. | 75 |
| 3. Retrieval baseline | Compare keyword retrieval with a semantic approach on the same questions. | 75 |
| 4. Grounded responses | Generate answers with citations to retrieved evidence. | 75 |
| 5. Evaluation and abstention | Measure retrieval coverage and reject unsupported answers. | 75 |
| 6. Release and maintenance | Document updates, deletion, ownership and regression checks. | 75 |
| Capstone | A working small RAG prototype with a ten-question evaluation, citation checks and an abstention policy. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Python and source-verification skills are required. Choose a local or existing authorized model environment.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-212 AI Agents & Workflow Orchestration
Build tool-using agent workflows with state, routing, approvals and recovery.
Learner study edition available · Enrollment closed
Open the complete learner study edition — lessons, practice quizzes, workbook, capstone and final assessment tasks. Instructor-reviewed enrollment remains closed.
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-115, NTL-117. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Represent an agent workflow as explicit states.
- Constrain tool use and approval boundaries.
- Recover from a tool failure without duplicating an action.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. State and task boundaries | Define states, allowed transitions and a terminal condition. | 75 |
| 2. Tool contracts | Specify inputs, output validation and permission scope for two tools. | 75 |
| 3. Routing and handoffs | Route three fictional requests to the correct bounded handler. | 75 |
| 4. Memory and provenance | Persist only required state and identify each source. | 75 |
| 5. Failure and recovery | Inject a timeout, retry safely and preserve an audit trail. | 75 |
| 6. Evaluation and handover | Test normal, denied and interrupted paths and document operation. | 75 |
| Capstone | An agent workflow using sandbox tools with explicit state, approvals, failure tests and logs. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
External sends and production writes stay disabled in course exercises.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-213 APIs for AI Integration
Connect AI services to applications and business systems through APIs.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-115, NTL-116. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Read and validate a request and response contract.
- Keep credentials out of source, logs and browser output.
- Implement bounded retries and idempotent action handling.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. HTTP and contracts | Inspect methods, status codes and a versioned JSON contract. | 75 |
| 2. Authentication boundaries | Load a test credential through an approved secret mechanism. | 75 |
| 3. Validation and errors | Reject missing fields, wrong types and unexpected response shapes. | 75 |
| 4. Timeouts and retries | Separate retryable failures from permanent failures. | 75 |
| 5. Webhooks and idempotency | Process a repeated signed test event only once. | 75 |
| 6. Integration documentation | Demonstrate a local API integration with redacted logs and a runbook. | 75 |
| Capstone | A tested API integration using a local mock service, with validation, failure handling and duplicate-event checks. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
No paid API is required for the assessed mock-service route.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-214 Applied Machine Learning for AI Integrators
Train and evaluate practical supervised models and compare them with foundation-model approaches.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-114, NTL-115. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Split data without leaking test information.
- Compare a simple supervised model with an appropriate baseline.
- Explain precision, recall and the effect of a decision threshold.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Problem and baseline | Choose a classification or regression target and a naive baseline. | 75 |
| 2. Training and evaluation splits | Create a split that respects time or group membership where needed. | 75 |
| 3. Features and preprocessing | Fit transformations on training data only. | 75 |
| 4. Model fitting | Train a small model and preserve environment and parameter details. | 75 |
| 5. Metrics and error analysis | Inspect confusion counts or residuals and analyze costly mistakes. | 75 |
| 6. Model selection and reporting | Compare approaches and defend a deployment or no-deployment recommendation. | 75 |
| Capstone | A reproducible experiment report with held-out results, a baseline and an error analysis. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Use a small licensed public or fictional dataset. Course outcomes do not depend on achieving a predetermined accuracy.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-215 Vector Databases & Semantic Search
Implement embeddings, vector retrieval, hybrid search and relevance testing.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-115, NTL-211. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Describe embedding similarity without treating it as truth.
- Build filtered semantic and hybrid retrieval.
- Measure ranking quality against labeled relevance judgments.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Embedding representations | Compare semantically related and unrelated query-document pairs. | 75 |
| 2. Index construction | Record embedding model, dimensions, IDs and metadata. | 75 |
| 3. Similarity and retrieval | Compare top-k results and investigate a misleading near neighbor. | 75 |
| 4. Metadata and access filters | Apply scope filters before a result reaches the answer stage. | 75 |
| 5. Hybrid search and reranking | Compare keyword, vector and combined rankings. | 75 |
| 6. Evaluation and maintenance | Measure retrieval performance and plan reindexing after a model change. | 75 |
| Capstone | A searchable collection with labeled queries, ranking comparisons and a reproducible index configuration. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
NTL-211 covers the end-to-end answer pipeline; this course concentrates on retrieval and indexing.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-216 AI Automation for Business Operations
Integrate AI into repeatable operational processes with controls and measurable outcomes.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-117, NTL-213. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Identify which steps should remain deterministic.
- Design approval and exception handling for business automation.
- Test duplicate, delayed and partial-failure events.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Process selection | Choose a repeatable process and define observable success. | 75 |
| 2. Deterministic and AI steps | Separate exact validation from judgment-based drafting. | 75 |
| 3. Routing and approvals | Create mutually exclusive paths and a human approval gate. | 75 |
| 4. Delivery and idempotency | Use a stable event identity and document replay behavior. | 75 |
| 5. Exceptions and recovery | Test a timeout, a malformed input and a failed downstream step. | 75 |
| 6. Operations and measurement | Prepare a runbook, ownership table and measured pilot report. | 75 |
| Capstone | A sandbox automation with a ten-case execution log and a documented manual recovery procedure. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
The existing AUT-301 written edition is a related source; it is not automatically a second credit.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-217 Cloud AI Deployment Foundations
Deploy and monitor AI-backed services using contemporary cloud patterns.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-213, NTL-116. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Package a small AI-backed service with separated configuration.
- Use health checks, logging and resource limits.
- Demonstrate rollback after a failed release.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Service boundaries | Specify the request path, dependencies and failure budget. | 75 |
| 2. Configuration and secrets | Separate environment configuration from code and logs. | 75 |
| 3. Packaging and staging | Build a reproducible service package and deploy to an authorized test environment. | 75 |
| 4. Observability | Record errors, latency and model or configuration versions. | 75 |
| 5. Release and rollback | Exercise a failed release and restore the prior version. | 75 |
| 6. Cost and handover | Estimate resource use and produce a tested operating runbook. | 75 |
| Capstone | A locally containerized or authorized staging deployment with health, failure and rollback evidence. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
A local container route is acceptable; purchasing cloud services is not required.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
Year 3 · Advanced specialization
Evaluate complex systems, defend architecture choices and produce evidence for release decisions.
NTL-311 Advanced Agentic AI Systems
Design multi-step autonomous and semi-autonomous systems with evaluation and human control.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-212, NTL-217. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Compare a single agent with a coordinated multi-agent design.
- Bound delegation, shared state and resource use.
- Evaluate failure recovery and human intervention across a complete task.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Architecture comparison | Explain when additional agents help and when they add failure paths. | 75 |
| 2. Coordination protocols | Define ownership, task receipts and completion criteria. | 75 |
| 3. Shared state and memory | Resolve conflicting updates without silently losing provenance. | 75 |
| 4. Permissions and budgets | Bound tool scopes, action counts and escalation rules. | 75 |
| 5. Adversarial evaluation | Test misleading handoffs, loops and partial failures in a sandbox. | 75 |
| 6. Architecture defense | Present measured results and justify the simplest adequate design. | 75 |
| Capstone | A bounded agent-system comparison with end-to-end traces, intervention tests and an architecture defense. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Advanced work must include observed results, not only an architecture diagram.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-312 AI Model Evaluation & Red Teaming
Develop rigorous evaluation suites for quality, safety, robustness and failure modes.
Learner study edition available · Enrollment closed
Open the complete learner study edition — lessons, practice quizzes, workbook, capstone and final assessment tasks. Instructor-reviewed enrollment remains closed.
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-211, NTL-214. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Design an evaluation set separate from development examples.
- Measure quality and failure severity with declared scoring rules.
- Make a release decision from observed evidence and uncertainty.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Evaluation specification | Define task, target users, success measures and failure severity. | 75 |
| 2. Dataset design | Create representative and challenge cases without train-test overlap. | 75 |
| 3. Scoring and calibration | Compare two reviewers on a shared rubric and resolve disagreements. | 75 |
| 4. Robustness testing | Test missing context, conflicting evidence and malicious source instructions. | 75 |
| 5. Regression analysis | Compare versions on the same fixed test set and report regressions. | 75 |
| 6. Release decision | Write a go, revise or hold decision with unresolved risks. | 75 |
| Capstone | An evaluation suite with at least twenty inputs, baseline and revision results, and a defended release memo. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Testing is restricted to owned or explicitly authorized systems and synthetic test data.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-313 AI Security & Adversarial Resilience
Analyze prompt injection, data leakage, tool abuse and defensive AI architecture.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-116, NTL-212. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Map trust boundaries in an AI workflow.
- Test injection, data leakage and tool misuse with harmless markers.
- Connect each mitigation to a repeatable regression test.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Threat modeling | Identify assets, entry points, trusted instructions and untrusted data. | 75 |
| 2. Prompt injection | Test document-borne instructions against a sandbox assistant. | 75 |
| 3. Information boundaries | Use synthetic canary strings to check unauthorized disclosure. | 75 |
| 4. Tool and retrieval controls | Enforce access scopes and validate tool arguments. | 75 |
| 5. Defense evaluation | Compare behavior before and after a control change. | 75 |
| 6. Incident response | Write containment, recovery and regression procedures. | 75 |
| Capstone | A threat model, an authorized twenty-case security test set and a mitigation report with reruns. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
No real credentials, third-party intrusion or attempts to bypass service safeguards are required.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-314 Enterprise RAG Architecture & Knowledge Engineering
Architect governed enterprise retrieval systems with provenance and access control.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-211, NTL-215, NTL-213. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Design retrieval with document-level access control.
- Preserve provenance through updates and deletion.
- Compare relevance, latency and operational cost under declared constraints.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Enterprise knowledge model | Define owners, document types and source-of-truth rules. | 75 |
| 2. Access-aware ingestion | Carry authorization metadata from source to indexed chunks. | 75 |
| 3. Retrieval architecture | Compare retrieval and reranking options against a realistic workload. | 75 |
| 4. Freshness and deletion | Test updates, withdrawal and stale-cache handling. | 75 |
| 5. Evaluation and observability | Separate unauthorized retrieval, bad ranking and unsupported generation. | 75 |
| 6. Architecture review | Defend a design using measured trade-offs and a maintenance plan. | 75 |
| Capstone | An enterprise retrieval design with a small working prototype, access tests and a freshness experiment. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Synthetic user roles and documents are sufficient; a production enterprise connection is unnecessary.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-315 Multimodal AI Systems
Integrate text, image, audio and video models into advanced applications.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-211, NTL-213, NTL-116. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Design a workflow spanning text and at least one other modality.
- Preserve source and timestamp alignment.
- Evaluate modality-specific uncertainty and accessibility.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Modality and task choice | Choose a bounded text-image or text-audio use case. | 75 |
| 2. Ingestion and consent | Document allowed sources, permissions and retention. | 75 |
| 3. Alignment and provenance | Keep transcript segments or image regions connected to evidence. | 75 |
| 4. Model orchestration | Validate inputs and outputs across two processing stages. | 75 |
| 5. Evaluation and fallback | Test noise, missing media and misleading content. | 75 |
| 6. Accessible delivery | Produce a usable output with alternatives and a failure disclosure. | 75 |
| Capstone | A multimodal prototype with source alignment, ten test cases, accessibility checks and a failure analysis. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
Use licensed or learner-created media and fictional identities.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
NTL-316 AI Systems Architecture & Integration Strategy
Evaluate enterprise AI architectures, integration trade-offs, observability and lifecycle risk.
Syllabus available · Teaching package in development
Planned instruction: 10 hours · 1 internal learning credit after verified completion.
Prerequisites: NTL-212, NTL-217, NTL-312. Equivalent experience may be reviewed by an instructor.
Learning outcomes
- Compare build, buy and hybrid integration options.
- Defend system boundaries and lifecycle ownership.
- Plan a staged release with measurable acceptance and rollback conditions.
Module sequence
| Module | Assessed activity | Minutes |
|---|---|---|
| 1. Business and system requirements | Translate a use case into measurable service and quality requirements. | 75 |
| 2. Architecture options | Compare three implementation approaches using declared criteria. | 75 |
| 3. Interfaces and data contracts | Specify ownership, failure paths and version boundaries. | 75 |
| 4. Governance and economics | Model review effort, operating cost and residual risk with explicit assumptions. | 75 |
| 5. Migration and release | Plan a reversible rollout with acceptance and rollback evidence. | 75 |
| 6. Decision defense | Present an architecture decision record and respond to a structured challenge. | 75 |
| Capstone | A complete architecture decision pack supported by a small integration proof and an evidence-based release plan. | 90 |
| Final assessment | Demonstrate the stated outcomes and explain your decisions. | 60 |
Tools and preparation
The final recommendation may be to delay or reject an AI component when the evidence does not support it.
Completion requirements
- Complete every module activity.
- Achieve at least 80 percent on module quizzes and the final assessment.
- Achieve at least 80 percent on the capstone rubric with all required evidence present.
- Obtain instructor verification of work and documented instructional hours.
Availability: Enrollment is not open. Start dates have not been announced.
Study and credit information
Each course plans six 75-minute modules, a 90-minute capstone and a 60-minute final assessment, totaling 600 instructional minutes. These are planned allocations. One internal Northline learning credit requires ten documented instructional hours and successful completion verified by an instructor.
Year levels describe curriculum progression. Internal learning credits are not represented as transferable university credit, external academic accreditation or professional licensure. This page describes syllabi and does not open paid enrollment.