LOCKEDIN / FORWARD-DEPLOYED ENGINEERING 101
Understand it.
Then build it.
You can be an experienced engineer and still be new to agents. Start with what each thing is and how it helps a person do useful work. Explore 58 original introductions across software, agents, cloud, security and delivery. Start here without a model key; each topic connects to deeper instruction and practice.
A place to start.
Choose a short reading route, or look up a concept below.
- 101UnderstandPlain language, visual walkthrough, vocabulary
- 102Build with guidanceOriginal lessons, examples and knowledge checks
- 103Practise and explainRunnable labs, failure cases and handover
Understand agents
Put the model inside an application that can execute permitted actions and check the result.
See the 8-concept route
For engineers who know software but are new to AI applications. Open the related definitions whenever a term is unfamiliar.
- 01The modelStart with what produces the text. Separate the model from the application using it.
- 02TokensSee the units a model reads and produces before reasoning about its input limits.
- 03ContextDecide what information reaches the model for this particular request.
- 04The agent loopYou are here
- 05WorkflowsSeparate a fixed sequence from a model-selected next action before choosing a framework.
- 06LangChainSee how a framework assembles the model, tools and application-controlled context.
- 07LangGraphMake the steps and shared state explicit when work needs to branch, wait or resume.
- 08MCPConnect capabilities through a shared protocol while keeping permission checks in the application and server.
Bookmark this page to return to this concept. The number shows your place in the reading route; it is not an assessment result.
An application built around an AI model
What is an AI agent?
An AI agent is software that uses an AI model to choose some of the steps needed to complete a task. It can call tools, inspect their results and decide what to do next within limits set by the application. The model is one component; the agent also needs code, tools and rules about when to stop.
An FDE needs to decide when model-driven choices help a real task and when a predictable sequence is enough. An agent is useful only if it improves the work and stays within its allowed scope.
UNDERSTAND THE SYSTEM
An agent chooses an action, then reads the result
An agent is an application that lets a model choose its next action within limits. The application runs tools; the model does not reach into your systems by itself.
Read the connections in plain English
- 01Your goal → Model
The goal and tool descriptions become part of the model's input.
- 02Model → Application checks
The model proposes lookup_order with order_id 104. A proposal is not permission.
- 03Application checks → Order lookup
Trusted code authorizes the caller and runs the lookup within its configured limits.
- 04Order lookup → Model
The tool result joins the context. The model can answer or propose another permitted action.
- 05Model → Answer or stop
The application checks the result and ends the loop; a budget limit can also force a stop.
Check your understanding: What changes if the model asks to cancel the order?
Nothing is cancelled by the model's words alone. A write tool would need to exist, pass authorization and satisfy the application's confirmation policy. A read-only agent should reject that action.
CONNECT IT TO WHAT YOU KNOW
Start with something familiar.
A normal program follows rules you wrote. An agent application can ask a model to choose the next permitted action based on what it observes, then use ordinary code to execute and check that action.
The model proposes; application code grants access, runs tools and enforces limits. A chatbot that only answers text does not necessarily have that action loop.
Words worth knowing
- Model / LLM
- The AI component that generates text or proposes an action; LLM means large language model.
- Tool
- A function the application makes available, such as looking up a procedure.
- Agent loop
- The repeated cycle of model choice, permitted tool use and checking whether to continue.
- Workflow
- The connected steps used to complete a task. Some or all can be fixed in advance.
Useful foundations
Open a definition and diagram here. Your lesson stays in place.
A trained statistical modelWhat is a large language model?
QUICK REFERENCE / 101
What is a large language model?
A large language model (LLM) is a trained neural network. Its learned numerical parameters and current input are used to calculate possible next tokens; a decoding process selects the continuation. A model can produce text, code or structured proposals. The application around it supplies context, decides what tools are available and controls real-world actions.
Choosing a model is one decision inside an application design, not the whole design.
A software contractWhat is an API?
QUICK REFERENCE / 101
What is an API?
An application programming interface (API) is an agreed way for one piece of software to use another. It specifies operations, inputs, outputs and errors. Some APIs are local functions; a web API exchanges requests and responses across a network.
An agent becomes useful through connections to business systems. Those connections need clear contracts and failure behavior.
ONE EXAMPLE / FICTIONAL CASE
Follow the connections.
For a maintenance question, a chatbot might write a general answer. An agent can decide to look up an approved procedure, read the result, then draft a response for a person to review. It cannot gain permission to send the response just by deciding to do so.
UNDERSTAND THE SYSTEM
An agent chooses an action, then reads the result
An agent is an application that lets a model choose its next action within limits. The application runs tools; the model does not reach into your systems by itself.
See what moves and why.
Read the connections in plain English
- 01Your goal → Model
The goal and tool descriptions become part of the model's input.
- 02Model → Application checks
The model proposes lookup_order with order_id 104. A proposal is not permission.
- 03Application checks → Order lookup
Trusted code authorizes the caller and runs the lookup within its configured limits.
- 04Order lookup → Model
The tool result joins the context. The model can answer or propose another permitted action.
- 05Model → Answer or stop
The application checks the result and ends the loop; a budget limit can also force a stop.
Check your understanding: What changes if the model asks to cancel the order?
Nothing is cancelled by the model's words alone. A write tool would need to exist, pass authorization and satisfy the application's confirmation policy. A read-only agent should reject that action.
Work through it, step by step.
Give it a task
A person asks for a draft reply to a maintenance request. The application supplies the request and the allowed actions.
What you have now: Task: prepare a draft; do not send it
Choose and use a tool
The model proposes a procedure lookup. Application code checks access, runs the lookup and returns its result.
What you have now: An approved procedure, or a clear refusal
Produce or stop
The model drafts from the returned evidence. If evidence is missing or the allowed steps are exhausted, the application stops or asks a person.
What you have now: A reviewable draft, or a request for help
FROM RECOGNITION TO REASONING
Separate a chatbot, a fixed workflow and an agent.
- 01
Describe a question answered with no tools.
- 02
Describe a fixed lookup-then-format sequence.
- 03
Describe one task where choosing the next diagnostic action adds value, and set a stopping rule.
Three examples with the decision owner marked.
Compare your reasoning when you are ready
If the next steps are already known, use a fixed workflow first. Introduce a model-controlled choice only where its flexibility helps the task and can be evaluated.
Keep notes in your own working copy. Moving through these cards does not record completion.
TURN THE IDEA INTO USEFUL WORK
You can now go deeper.
A chatbot answers. A fixed workflow follows prescribed steps. An agent can choose some steps. None of these names grants permission to act.
Academy lessons use your existing training access. Free guides stay open.
Learn with the original creators
Compare the discussion of simple workflows and agents with your own task. Return with one reason to use an agent and one reason not to.