When does class start/end?
Class hours may vary, please reach out to contact@ascendientlearning.com if you have any questions.
Foundations of AI Architecture
Design AI solution architectures from use-case framing through Go/No-Go recommendation. Use-case framing separates architecturally significant choices from routine integration, then contrasts an AI...
Read More
Duration
3 days
Course Code
GAI-1701
Available Formats
Classroom
Overview
Course Description
Design AI solution architectures from use-case framing through Go/No-Go recommendation. Use-case framing separates architecturally significant choices from routine integration, then contrasts an AI architecture pattern menu — retrieval-grounded, generative, agentic, tool-use, classification-and-extraction, fine-tuned, and hybrid — so the architect picks deliberately. The data layer and non-functional requirements span source-of-truth alignment, retrieval and embedding stores, lineage, freshness, access control, latency, accuracy, TCO and per-token economics, and qualitative criteria including explainability, auditability, and safety. Sourcing and reference-architecture critique walk make-buy-borrow-and-contract trade-offs across foundation models, hosted APIs, and open stacks, contracting clauses for tenancy and exit, and adapting vendor diagrams to cloud, on-prem, or hybrid deployment constraints. Communicating decisions and recommending Go/No-Go/Pivot draw on ADRs, executive summaries, stakeholder-tuned visuals, and the synthesis move that backs a recommendation with named risks and validation steps. Hands-on labs produce pattern-fit critiques, NFR worksheets, sourcing trade-off matrices, ADRs, and a capstone Go/No-Go recommendation. The course is designed for solution, enterprise, and AI architects plus technical leads taking on AI scope for the first time.
Skills Gained
By the end of this course, participants will be able to:
- Frame AI use cases by separating architecturally significant choices from routine integration decisions
- Select an AI architecture pattern from a documented menu that matches use-case constraints and trade-offs
- Specify non-functional requirements, total cost of ownership, plus evaluation criteria for an AI workload
- Compare make, buy, borrow, and contract sourcing options against a documented decision rubric
- Draft an architecture decision record that defends an AI investment to executive and technical audiences
- Recommend Go, No-Go, or Pivot on an AI proposal supported by named risks and validation steps
Who Can Benefit
This course is designed for:
- Solution Architects
- Enterprise Architects
- AI Architects
- Technical Leads
Prerequisites
Participants should enter this course with:
- Familiarity with enterprise software architecture
- General awareness of AI and machine-learning concepts
Organizational Objectives
This course assists organizations to:
- Reduce stalled AI proposals through a shared architecture-decision framework that converges teams faster
- Lower vendor-lock-in risk by applying make-buy-borrow rubrics plus exit-clause review during sourcing
- Improve executive confidence in AI investments through ADR-backed Go/No-Go recommendations
- Build architecture capability across the enterprise-architecture function for AI-bearing initiatives
Software
All attendees must have a modern web browser and an Internet connection.
Course Details
Course Details
Framing AI Use Cases and Architectural Concerns
By the end of this module, you will be able to separate architecturally significant choices from routine integration decisions, identify the recurring decision points that AI introduces, and draft a use-case frame that prepares a workload for pattern selection and NFR work downstream.
- Architecturally significant vs routine integration decisions
- Where AI changes the architecture versus where it is just another service
- Recurring decision points — data freshness, evaluation method, sourcing, autonomy
- Stakeholder framing — business outcome, user surface, risk surface
- Use-case canvas — inputs, outputs, success metric, failure mode
- Anti-frames — AI-as-feature-checkbox, model-first thinking, vendor-led scoping
- <b>Hands-on Lab:</b> Apply the use-case canvas to a candidate AI workload and produce a one-page frame that names its architecturally significant choices and the decisions deferred to later modules.
Choosing AI Architecture Patterns
By the end of this module, you will be able to navigate the AI architecture pattern menu, distinguish patterns that look similar but carry different architectural commitments, and select a pattern that matches your use case’s constraints rather than its vocabulary.
- The AI architecture pattern menu — retrieval-grounded, generative, agentic, tool-use, classification-and-extraction, fine-tuned, hybrid
- Tool-use and function-calling as distinct from full agency
- Classification and extraction as a first-class non-generative pattern
- Pattern composition — most production systems combine two or three patterns
- LoRA, PEFT, full fine-tuning, and RAG as a decision branch, not a single pattern
- Common mismatches — agentic-where-pipeline-fits, fine-tune-where-RAG-fits
- <b>Hands-on Lab:</b> Critique three candidate workloads against the pattern menu and produce a pattern-fit recommendation for each that names its primary pattern, composed patterns, and rejected alternatives.
Architecting the Data Layer for AI
By the end of this module, you will be able to align an AI workload’s data layer with the source of truth, design retrieval and embedding infrastructure that respects access control and freshness, and name the lineage and metadata commitments the workload will live with for years.
- Source-of-truth alignment and the cost of bypassing it
- Retrieval and embedding stores — scope, refresh cadence, freshness budgets
- Lineage and metadata — document IDs, chunking strategy, version pins
- Access control on retrieval — per-document ACLs, identity-aware retrieval
- Multimodal data — text, image, audio, video as retrieval surfaces
- Data-quality gates upstream of the model
- <b>Hands-on Lab:</b> Design the data layer for a candidate workload through a worksheet that names sources, retrieval and embedding stores, freshness budgets, ACL sources, and lineage metadata.
Setting Non-Functional Requirements and Evaluation Criteria
By the end of this module, you will be able to specify the non-functional requirements an AI workload must hit, name the cost levers that keep them affordable, and define the evaluation criteria that prove the workload is meeting them in production.
- Latency budgets, accuracy targets, and cost ceilings as paired numbers
- Total cost of ownership and per-token economics
- Cost levers — model-router cascades, prompt and semantic caching, batching, output-token discipline
- Qualitative criteria — explainability, auditability, safety, fairness
- Evaluation criteria as architectural commitments, not afterthoughts
- Tracing NFRs to NIST AI RMF, ISO/IEC 42001, and EU AI Act obligations
- <b>Hands-on Lab:</b> Draft an NFR worksheet for a candidate workload that pairs each requirement with a measurement method, a named cost lever, and the framework obligation it traces to.
Sourcing AI - Make, Buy, Borrow, and Contracting
By the end of this module, you will be able to compare make, buy, borrow, and contract sourcing options for an AI workload, identify the contract clauses that protect tenancy and exit, and produce the initial sourcing decision that downstream operations will build on.
- Make, buy, borrow, contract trade-offs across foundation models, hosted APIs, and open stacks
- Tenancy models — shared API, dedicated capacity, single-tenant inference, BYO-key
- Contract clauses — training-data opt-out, IP indemnification, data residency, deprecation notice, exit and portability
- Cloud, on-prem, and hybrid sourcing implications
- Sub-processor lists, AUP alignment, and audit-log support
- Boundary with production operationalisation (handed to GAI-2703)
- <b>Hands-on Lab:</b> Build a sourcing trade-off matrix for a candidate workload that ranks options across cost, control, exit risk, and contract maturity, and recommend an initial sourcing decision.
Critiquing and Adapting Reference Architectures
By the end of this module, you will be able to read a vendor reference architecture critically, name what it optimises for and what it hides, and adapt its diagram to the constraints of a specific workload without inheriting its assumptions.
- Reading hyperscaler, data-platform, and GPU-vendor reference architectures
- What a reference architecture optimises for and what it omits
- Adapting a vendor diagram to specific constraints — tenancy, data residency, compliance, cost
- Recognising vendor lock-in baked into the picture
- Scenario critique as a teachable architect skill
- When to walk away from the reference and design fresh
- <b>Hands-on Lab:</b> Critique two vendor reference architectures against a candidate workload and produce an adapted architecture diagram with annotations naming the changes and the constraints driving them.
Communicating AI Architecture Decisions
By the end of this module, you will be able to draft an architecture decision record that defends an AI investment to both executive and technical audiences, produce stakeholder-tuned visuals, and avoid the framing traps that erode executive trust in AI proposals.
- Architecture decision records for AI — context, options, decision, consequences, validation plan
- Executive summary structure — outcome, investment, risk, validation milestones
- Stakeholder-tuned visuals — same architecture for engineering, security, and finance
- Naming risks honestly without losing the room
- Pre-empting common executive objections to AI proposals
- The seam between ADR and the Go/No-Go recommendation
- <b>Hands-on Lab:</b> Draft an ADR plus a one-page executive summary for a candidate AI workload, paired with two stakeholder-tuned diagrams: one for engineering and one for the executive sponsor.
Recommending Go, No-Go, or Pivot
By the end of this module, you will be able to synthesise pattern, NFR, sourcing, and reference-architecture work into a Go, No-Go, or Pivot recommendation backed by named risks and validation steps, and defend the recommendation in front of an executive audience.
- Synthesising the architecture work into one recommendation
- Go, No-Go, Pivot as three distinct artefacts with different downstream commitments
- Naming risks and validation steps that gate each option
- Distinguishing pivot from no-go — changing scope versus stopping
- Defending the recommendation under executive challenge
- Tracking the recommendation into the next phase of the workload
- <b>Hands-on Lab:</b> Produce a capstone Go/No-Go/Pivot recommendation on a candidate AI workload that integrates the prior labs into a single defensible artefact with named risks and validation steps.
Schedule
FAQ
Does the course schedule include a Lunchbreak?
Classes typically include a 1-hour lunch break around midday. However, the exact break times and duration can vary depending on the specific class. Your instructor will provide detailed information at the start of the course.
What languages are used to deliver training?
Most courses are conducted in English, unless otherwise specified. Some courses will have the word "FRENCH" marked in red beside the scheduled date(s) indicating the language of instruction.
What does GTR stand for?
GTR stands for Guaranteed to Run; if you see a course with this status, it means this event is confirmed to run. View our GTR page to see our full list of Guaranteed to Run courses.
How do I find an Ascendient Learning training location?
We have training locations across the United States and Canada - View a complete list of classroom training locations.
Which delivery formats are available?
At Ascendient Learning, we offer training that is Instructor-Led, Online, Virtual, and Self-Paced.
Does Ascendient Learning deliver group training?
Yes, we provide training for groups, individuals and private on sites. View our group training page for more information.
What does vendor-authorized training mean?
As a vendor-authorized training partner, we offer a curriculum that our partners have vetted. We use the same course materials and facilitate the same labs as our vendor-delivered training. These courses are considered the gold standard and, as such, are priced accordingly.
Is the training too basic, or will you go deep into technology?
It depends on your requirements, your role in your company, and your depth of knowledge. The good news about many of our learning paths, you can start from the fundamentals to highly specialized training.
How up-to-date are your courses and support materials?
We continuously work with our vendors to evaluate and refresh course material to reflect the latest training courses and best practices.
Are your instructors seasoned trainers who have deep knowledge of the training topic?
Ascendient Learning instructors have an average of 27 years of practical IT experience and have also served as consultants for an average of 15 years. To stay current, instructors spend at least 25 percent of their time learning new, emerging technologies and courses.
Do you provide hands-on training and exercises in an actual lab environment?
Lab access is dependent on the vendor and the type of training you sign up for. However, many of our top vendors will provide lab access to students to test and practice. The course description will specify lab access.
Will you customize the training for our company’s specific needs and goals?
We will work with you to identify training needs and areas of growth. We offer a variety of training methods, such as private group training, on-site of your choice, and virtually. We provide courses and certifications that are aligned with your business goals.
How do I get started with certification?
Getting started on a certification pathway depends on your goals and the vendor you choose to get certified in. Many vendors offer entry-level IT certification to advanced IT certification that can boost your career. To get access to certification vouchers and discounts, please contact info@ascendientlearning.com.
Will I get access to content after I complete a course?
You will get access to the PDF of course books and guides, but access to the recording and slides will depend on the vendor and type of training you receive.
How do I request a W9 for Ascendient Learning?
View our filing status and how to request a W9.
Reviews
very good and spcecific course and above all a very good instructor. In few days I have learned a lot.
Great class I learned a great deal from the material. There would seem to a large amount that I need to learn about.
ExitCertified provided us with a great opportunity to learn more about React and in easy to follow way.
Brandon was a great instructor. The virtual course materials and labs provided were very informative.
The exit certified aws course provided a good introduction to the tools available on aws.
