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
See the units a model reads and produces before reasoning about its input limits.
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.
- 02TokensYou are here
- 03ContextDecide what information reaches the model for this particular request.
- 04The agent loopPut the model inside an application that can execute permitted actions and check the result.
- 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.
Units of model input and output
What are tokens and token budgets?
Tokens are the units a model uses to represent content. A token can be a word, part of a word or another encoded element. Token counts affect context capacity, output limits, latency and often price. Characters, words and tokens are not interchangeable counts.
Unbounded loops and large tool responses can exhaust cost and capacity before useful work finishes.
UNDERSTAND THE SYSTEM
Tokens consume both context space and budget
A token is a unit used to encode model input or output. Token counts depend on the model's tokenizer; there is no universal one-word-to-one-token rule.
Read the connections in plain English
- 01Instructions + data → Model tokenizer
Count using the selected model's supported tokenizer or usage tooling.
- 02Model tokenizer → Context budget
Check input size before sending an oversized request.
- 03Context budget → Output allowance
Keep sufficient room within the selected model's rules.
- 04Output allowance → Recorded usage
Record actual usage rather than treating an estimate as a bill.
Check your understanding: Will the same text always cost the same across models?
No. Tokenization, pricing categories, caching and generated output can differ. Compare actual task-level usage with current provider terms.
CONNECT IT TO WHAT YOU KNOW
Start with something familiar.
Treat a request like a bounded packet with room for both the question and the answer. Filling every available space with input leaves no sensible output allowance.
Budget the complete loop: instructions, history, tool descriptions, retrieved evidence, output and repeated calls. A limit on one response is not a spending cap for an entire agent task.
Words worth knowing
- Context window
- The model's capacity for a request's context, subject to provider rules.
- Output limit
- A maximum on generated tokens for a response.
- Usage
- Provider-reported consumption for the actual request.
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.
ONE EXAMPLE / FICTIONAL CASE
Follow the connections.
The sentence “The pump needs inspection.” has 26 Unicode code points but five tokens in the lab’s cl100k_base encoding. The leading space and pump share one token.
UNDERSTAND THE SYSTEM
Tokens consume both context space and budget
A token is a unit used to encode model input or output. Token counts depend on the model's tokenizer; there is no universal one-word-to-one-token rule.
See what moves and why.
Read the connections in plain English
- 01Instructions + data → Model tokenizer
Count using the selected model's supported tokenizer or usage tooling.
- 02Model tokenizer → Context budget
Check input size before sending an oversized request.
- 03Context budget → Output allowance
Keep sufficient room within the selected model's rules.
- 04Output allowance → Recorded usage
Record actual usage rather than treating an estimate as a bill.
Check your understanding: Will the same text always cost the same across models?
No. Tokenization, pricing categories, caching and generated output can differ. Compare actual task-level usage with current provider terms.
Work through it, step by step.
Count input
Use the target model's tokenizer or documented counting method; do not use a universal words-to-tokens rule.
What you have now: An input estimate
Reserve output
Allocate a useful output limit and include the total across concurrent requests.
What you have now: A complete capacity request
Bound the task
Enforce per-task calls, elapsed time and spending limits in the host, including retries.
What you have now: An enforceable budget
FROM RECOGNITION TO REASONING
Inspect a sentence before budgeting the request.
- 01
Open the free token lab and compare its sentence, code and Japanese samples.
- 02
Count 5 text tokens + 24 other input + 80 history + 48 output allowance.
- 03
Compare that reservation with the hypothetical 128-token window; explain what to change.
An observed token representation and a budget with assumptions named.
Compare your reasoning when you are ready
The reservation is 157, or 29 over this teaching window. Tokenization converts text to IDs; it does not generate an answer. The local encoding is not a Claude counter or a bill.
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.
Budget the complete loop: instructions, history, tool descriptions, retrieved evidence, output and repeated calls. A limit on one response is not a spending cap for an entire agent task.
Academy lessons use your existing training access. Free guides stay open.