VDF AI API and Integration Engineering
VDF AI API and Integration Engineering is an instructor-led course for developers who connect VDF AI to their own systems. Over four live half-day sessions you authenticate, run agents and networks, manage data connections and vector search, and embed agents in your product, finishing with a capstone integration that earns the VDF AI Certified Developer certificate.
- 4 live half-days
- 6 modules + capstone
- Remote or on-site
- VDF AI Certified Developer
- Level
- Advanced
- Format
- Live and instructor-led, remote or on-site
- Length
- Four live half-day sessions (3.5 hours each)
- Audience
- Software engineers and integration developers who build on VDF AI
- Cost
- Free for customers and partners; quoted for other teams
- Certificate
- VDF AI Certified Developer
- Reply to applications
- Within 2 business days
- Labs
- One after every module
What you will be able to do
- Authenticate once and call every published VDF AI API with the same token
- Handle pagination, errors and long-running requests without fragile polling code
- Run agents and networks from your own services and read back their evidence
- Give agents governed access to your databases through the Data API
- Serve the VDF AI tool catalogue to external MCP clients and embed agents in your product
Prerequisites
- Comfortable calling REST APIs from a language of your choice
- Access to a VDF AI deployment for the labs
- Recommended free path: Build Enterprise AI Agents
6 modules and a capstone
The modules run across the four sessions. Each ends with a lab your team completes in a VDF AI environment.
-
Authentication, requests and errors
The conventions every VDF AI API shares, so the rest of the course is about what you build rather than how to call it.
- Signing in and the access token every VDF AI API accepts
- Request conventions, pagination and the error model
- Long-running requests: starting work and collecting the result
- Base paths for the Chat, Agent Hub, Networks and Data APIs
- Lab
- Build a small API client that authenticates, pages through your agents and recovers cleanly from a deliberately failing request.
- Outcome
- A reusable client with token handling, pagination and retries that the later labs build on.
-
Conversations, documents and integrations
The Chat API: where users, documents and connected systems enter the platform.
- Chat sessions and their messages
- Uploading documents to a knowledge base
- Authorising and synchronising integrations such as Jira, Confluence and GitHub
- Network triggers that start a network when something happens in a connected system
- Lab
- Upload a document set, synchronise one integration and ask questions over both in a chat session driven from code.
- Outcome
- Programmatic control of the knowledge your agents search.
-
Running agents with the Agent Hub API
Agents, the tools they may use and the executions that record what they did.
- Agents, workspaces and domains
- Tools: built-in tools, HTTP tools you define and remote MCP servers
- Skills and their versions
- Executions, sessions and sharing
- Lab
- Define an HTTP tool for an internal service, grant it to an agent and run the agent from your client, reading the execution record.
- Outcome
- Agents that call your own services, started and inspected from code.
-
Orchestrating networks with the Networks API
Multi-agent networks as directed graphs you can create, run and audit through the API.
- Creating and executing networks
- Following runs and reading the evidence behind each result
- Approvals and artifacts
- Intent endpoints and templates that turn a plain-language task into a network
- Lab
- Create a network from a template, run it, act on its approval step and collect the artifacts it produces.
- Outcome
- Multi-agent processes embedded in your own systems, with their records intact.
-
Governed data access with the Data API
Letting agents work with databases where they live, without handing them credentials.
- Registering database connections
- Exploration and profiling
- Vector indexes and semantic search over tables
- Generating fine-tuning datasets
- Lab
- Register a database connection, build a vector index over one table and run semantic search from code.
- Outcome
- Database-backed retrieval that stays inside the platform’s governance.
-
Embedding and extending VDF AI
Putting agents in front of your own users and your tools in front of other clients.
- Serving the tool catalogue to external MCP clients through the MCP endpoint
- Embedding agents in your own product
- Escalating agentic flows to the API from workflow tools
- Least-privilege tokens and keeping integrations auditable
- Lab
- Connect an external MCP client to the VDF AI tool catalogue and embed an agent in a sample web application.
- Outcome
- A pattern for shipping VDF AI capabilities inside your own software.
Capstone: an integration from your own organisation
Build a working integration for a real use case, for example a support flow that synchronises tickets, runs a network and writes the result back, then walk a VDF AI engineer through its design, error handling and audit trail.
VDF AI Certified Developer
Awarded to developers who complete VDF AI API and Integration Engineering and pass the capstone review.
- Authenticating and calling the published VDF AI APIs
- Running and auditing agents and networks from code
- Governed database access and vector search through the Data API
- Embedding agents and exposing tools over MCP
What your team gets
- Four live half-day sessions (3.5 hours each)
- A hands-on lab after every module
- A materials pack: session slides and lab guides
- A capstone review with a VDF AI engineer
- The VDF AI Certified Developer certificate on passing the capstone
Apply, agree dates, learn
- Send the application below. It takes two minutes.
- We reply within 2 business days to confirm eligibility and agree dates.
- Your team receives the materials pack, then joins four live half-day sessions.
Also available: Production Agentic Systems: Multi-Agent, RAG and Governance.