“The Microsoft Fabric implementation program gave our data engineering team a structured path from legacy pipelines to a modern lakehouse architecture.”
Enterprise Program Brief
Applied AI Engineering with Microsoft Foundry
Build a grounded enterprise assistant with bounded tools and measurable quality. Proposed instructor-led duration; delivery subject to environment and lab validation.
VNode ITeSPractitioner / specialist
Download Program Brief (PDF)Program Snapshot
Engagement scoped to cohort
Duration56 hours
FormatPrivate cohort
LevelPractitioner / specialist
Course CodeE02
- MCT-led delivery
- Private cohort / virtual ILT
- Typical response: 1 business day
Scoping call typically scheduled within one business day
VNode ITeS delivers Applied AI Engineering with Microsoft Foundry as an MCT-led VNode ITeS program for Python developers, AI engineers. Typical duration is 56 hours (Private cohort). Delivery stays scoped to how your teams will actually use the platform after the room empties. Private cohorts can shift emphasis by role mix, workspace or repo constraints, and rollout timing.
Audience Profile
Built for these roles
Python developers, AI engineers
- Tailored Team Delivery
Overview
Executive overview
Build a grounded enterprise assistant with bounded tools and measurable quality.
Environment and licensing
Isolated Azure subscription/resource group Scoped deployment rights and resource-provider access Approved region, quota and optional model deployment Cost budget and cleanup plan Delivery depends on environment and lab validation. This outline does not include step-by-step lab instructions.
Prerequisites
- Python API client competence
- Azure identity and resource fundamentals
- AI and Generative AI Foundations, or equivalent demonstrable experience
- Python, APIs and Git for AI Engineering, or equivalent demonstrable experience
- Azure Foundations for AI Delivery, or equivalent demonstrable experience
Program Outcomes
Capabilities your teams will gain
- Deliver a grounded Foundry application with a bounded tool
- Compare versions using schema, retrieval and task-quality evidence
- Prove denied access, safe action approval and injection handling
- Provide reproducible deployment, tracing and operational ownership
Enterprise Delivery
Private cohort delivery
VNode-designed, vendor-aligned instructor-led program. Scope role mix, delivery format, and rollout goals for your cohort.
Curriculum
Curriculum roadmap
Proposed instructor-led duration; delivery subject to environment and lab validation.
- Application and model contracts
- Grounding and retrieval engineering
- Prompt and hosted agent choices
- Tools, APIs and controlled execution
- Evaluation and adversarial testing
- Deployment and observability
1Module 1 · 8 hours
Application and model contracts
+
Module 1 · 8 hours
Application and model contracts
Build a model client with validated response structure and compare approved deployments.
Business acceptance
- Define input, response and action contracts
- Choose measurable business acceptance criteria
Model selection
- Compare suitable model capabilities and deployment constraints
- Use a fixed representative comparison set
Structured output
- Validate structured fields against a schema
- Handle incomplete or invalid model output
Cost and latency baseline
- Measure baseline latency and usage
- Record model configuration and cost assumptions
Concept and design 4h · applied exercise 3h · review 1h. Included within module duration; no extra hours. Refine during delivery planning without silently changing published course hours.
Applied exercise: Application and model contracts
Build a model client with validated response structure and compare approved deployments.
Expected output. Model contract and baseline results
Acceptance. Baseline outputs meet the declared schema and model selection is evidence-based.
2Module 2 · 8 hours
Grounding and retrieval engineering
+
Module 2 · 8 hours
Grounding and retrieval engineering
Build a permission-aware retrieval path over synthetic policy documents.
Document ingestion
- Prepare documents with source identifiers and versions
- Define refresh and withdrawal requirements
Chunking
- Compare chunk size, overlap and metadata choices
- Avoid splitting information needed for a coherent answer
Hybrid retrieval concepts
- Compare keyword, vector and hybrid retrieval concepts
- Test relevance with reference questions
Citations and access filtering
- Validate citations against retrieved evidence
- Enforce access filtering before generation
Concept and design 4h · applied exercise 3h · review 1h. Included within module duration; no extra hours. Refine during delivery planning without silently changing published course hours.
Applied exercise: Grounding and retrieval engineering
Build a permission-aware retrieval path over synthetic policy documents.
Expected output. Retrieval pipeline and relevance tests
Acceptance. The retrieval path returns relevant permitted evidence and handles unsupported questions explicitly.
3Module 3 · 8 hours
Prompt and hosted agent choices
+
Module 3 · 8 hours
Prompt and hosted agent choices
Implement one bounded agent and justify its runtime and framework choice.
Managed versus custom orchestration
- Compare configured prompt agents with hosted custom code
- Choose the simplest runtime meeting requirements
Agent Framework
- Use a tested Agent Framework provider where suitable
- Document package and API support boundaries
State
- Define session state and conversation ownership
- Distinguish temporary context from durable memory
Tool contracts
- Specify bounded tool contracts and termination conditions
- Avoid unneeded multi-agent decomposition
Concept and design 4h · applied exercise 3h · review 1h. Included within module duration; no extra hours. Refine during delivery planning without silently changing published course hours.
Applied exercise: Prompt and hosted agent choices
Implement one bounded agent and justify its runtime and framework choice.
Expected output. Agent implementation and architecture decision
Acceptance. The runtime decision is justified and the bounded agent terminates correctly in tested cases.
4Module 4 · 8 hours
Tools, APIs and controlled execution
+
Module 4 · 8 hours
Tools, APIs and controlled execution
Connect an internal request API with strict input validation and explicit action confirmation.
Function and MCP concepts
- Define function and MCP tool interfaces
- Validate arguments before any action
Authentication
- Use resource-side authentication and authorisation
- Avoid trusting model-generated identity claims
Approval
- Require approval for consequential actions
- Separate read-only from state-changing operations
Retries and idempotency
- Set timeouts, retries and idempotency keys
- Reject duplicated or invalid operations
Concept and design 4h · applied exercise 3h · review 1h. Included within module duration; no extra hours. Refine during delivery planning without silently changing published course hours.
Applied exercise: Tools, APIs and controlled execution
Connect an internal request API with strict input validation and explicit action confirmation.
Expected output. Controlled tool integration
Acceptance. The tool integration denies unapproved actions and handles retry without duplicate business effects.
5Module 5 · 8 hours
Evaluation and adversarial testing
+
Module 5 · 8 hours
Evaluation and adversarial testing
Compare a baseline and revised version using deterministic checks and calibrated judges.
Gold datasets
- Construct representative normal and edge-case datasets
- Separate development examples from holdout checks
Groundedness
- Measure groundedness and reference correctness
- Calibrate judge scores with human review
Task success
- Evaluate tool selection and end-to-end task completion
- Inspect failures beyond response fluency
Injection and refusal tests
- Test malicious source instructions and data-exposure attempts
- Define refusal and safe-escalation acceptance
Concept and design 4h · applied exercise 3h · review 1h. Included within module duration; no extra hours. Refine during delivery planning without silently changing published course hours.
Applied exercise: Evaluation and adversarial testing
Compare a baseline and revised version using deterministic checks and calibrated judges.
Expected output. Evaluation suite and comparison report
Acceptance. A revised version is compared on the same dataset with recorded regressions and residual risks.
6Module 6 · 8 hours
Deployment and observability
+
Module 6 · 8 hours
Deployment and observability
Deploy to the approved environment and trace an injected tool failure.
Scoped identity
- Deploy with a scoped workload identity
- Keep secrets outside code and traces
Secrets
- Configure explicit environment and model references
- Record reproducible package and application versions
Tracing
- Inspect request, retrieval and tool traces
- Minimise sensitive telemetry
Release and rollback concepts
- Define release, smoke-test and rollback decisions
- Check cost and operational ownership
Concept and design 4h · applied exercise 3h · review 1h. Included within module duration; no extra hours. Refine during delivery planning without silently changing published course hours.
Applied exercise: Deployment and observability
Deploy to the approved environment and trace an injected tool failure.
Expected output. Deployment and diagnostic evidence
Acceptance. The deployed application can be traced, diagnosed and returned to a known configuration.
7Module 7 · 8 hours
Applied engineering capstone
+
Module 7 · 8 hours
Applied engineering capstone
Deliver the assistant with quality results, access tests and a reproducible handover.
End-to-end acceptance
- Complete the agreed user journey against reference cases
- Demonstrate answers and controlled actions
Access boundaries
- Show denied-access and injected-failure behaviour
- Explain unresolved threat and reliability boundaries
Operational documentation
- Present a release and operating decision
- Include reproducible configuration and quality evidence
Demonstration
- Defend architecture and model trade-offs
- Hand over support, ownership and change assumptions
Concept and design 2h · applied exercise 4h · review 2h. Included within module duration; no extra hours. Refine during delivery planning without silently changing published course hours.
Capstone: Applied engineering capstone
Deliver the assistant with quality results, access tests and a reproducible handover.
Expected output. Assessed Foundry AI application
Acceptance. The capstone supplies a working application plus quality, access and operational evidence.
Capstone
Deliver the assistant with quality results, access tests and a reproducible handover. Assessed Foundry AI application Evidence of relevant access, quality and risk controls Demonstration and limitations record
Assessment
Scenario-based practical assessment Deliver a grounded Foundry application with a bounded tool 30% Compare versions using schema, retrieval and task-quality evidence 25% Prove denied access, safe action approval and injection handling 25% Provide reproducible deployment, tracing and operational ownership 20% Proposed VNode outline-level assessment; no vendor accreditation or completed lab validation implied.
Participant deliverables
- Assessed Foundry AI application
- Evidence of relevant access, quality and risk controls
- Demonstration and limitations record
Scope boundaries
Step-by-step lab instructions; Runnable solution code; Packaged source datasets; Executed environment validation
Delivery options
Count teaching hours, not calendar days. Exclude breaks. Split into 4-hour or 8-hour teaching sessions; schedule complex project work in contiguous blocks.
Delivery Models
Delivery models
Virtual ILTOnsiteHybridExecutive WorkshopBootcampWeekend
Engagement Fit
Engagement fit
- Private cohort delivery
- Practitioner / specialist practitioner depth
- Business outcome alignment
Enterprise Customization
Enterprise customization
Tailor this program to your organization's priorities: Build a grounded enterprise assistant with bounded tools and measurable quality.
- Private cohort scheduling
- Equivalent-experience entry review
Resources
Program resources
Yes. Most enterprise clients prefer private delivery scoped to role mix, timezone, and rollout timeline. We align lab environments and scenarios to your tenant context where applicable.
Enterprise Proof
Trusted delivery outcomes
“We needed a partner who understood both the technical depth of Azure OpenAI and the governance requirements of an enterprise.”
Banking & Finance
Representative Enterprise Banking Team
The focus was not just on tooling knowledge, but on helping teams work from a shared operating model as they adopted a more modern data platform.
- Clearer platform operating model across teams
- Improved confidence in modern data stack adoption
Healthcare
Representative Healthcare Product Team
The engagement helped product and engineering stakeholders move from interest in AI to clearer implementation choices, security expectations, and prototyping discipline.
- Stronger alignment between product and engineering teams
- Improved clarity on prototype-to-production requirements
Delivery Capability
Enterprise-grade instruction
MCT-led delivery
Programs led by Microsoft Certified Trainer practitioners
Enterprise program oversight
Founder-led specialist delivery with structured rollout planning
Global delivery
APAC · EMEA · Americas · Virtual & Onsite
Implementation-focused
Hands-on labs aligned to production scenarios
