Prerequisites Basic computer literacy and a willingness to document experiments.
Learning objectives Explain the stage-specific model and its limits Apply the listed concepts using a controlled exercise Record results, failure cases and a reproducible evidence trail
Topics to learn Task, audience and success criteria Context, constraints and output format Instruction hierarchy and examples Prompt templates and variable boundaries Prompt Engineering specialization: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Tools Prompt version notes JSON examples Evaluation spreadsheet
Hands-on exercise Write prompts for three different task types and define pass/fail examples. Context for this route: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Suggested project A prompt specification with explicit inputs, outputs and acceptance criteria. The outcome must directly demonstrate Prompt Engineering, not merely list tools.
Learning resources Completion criteria A reviewer can judge outputs against objective criteria. Evidence must be reproducible and tied to Prompt Engineering.
Next: Few-shot examples and structured generation
✓ Mark this node as complete
Prerequisites Completion of the previous stage and a working local practice environment.
Learning objectives Explain the stage-specific model and its limits Apply the listed concepts using a controlled exercise Record results, failure cases and a reproducible evidence trail
Topics to learn Zero-shot vs few-shot trade-offs Example diversity and bias JSON schema and parsing Refusal and invalid-output handling Prompt Engineering specialization: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Tools JSON Schema Pydantic LLM API client
Hands-on exercise Create a structured extraction prompt and test incomplete or malformed inputs. Context for this route: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Suggested project A validated extraction workflow with a small labelled test set. The outcome must directly demonstrate Prompt Engineering, not merely list tools.
Learning resources Completion criteria Outputs either satisfy the schema or fail safely. Evidence must be reproducible and tied to Prompt Engineering.
Next: Reasoning support and task decomposition
✓ Mark this node as complete
Prerequisites Completion of the previous stage and a working local practice environment.
Learning objectives Explain the stage-specific model and its limits Apply the listed concepts using a controlled exercise Record results, failure cases and a reproducible evidence trail
Topics to learn Decomposition into verifiable subtasks Context ordering and summarization Grounding and source attribution When not to expose or rely on hidden reasoning Prompt Engineering specialization: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Tools Prompt templates Retrieval test set Unit tests
Hands-on exercise Build a staged synthesis workflow that produces source-linked conclusions. Context for this route: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Suggested project A multi-step prompt chain with explicit intermediate data contracts. The outcome must directly demonstrate Prompt Engineering, not merely list tools.
Learning resources Completion criteria Each result is grounded in supplied evidence and checked before use. Evidence must be reproducible and tied to Prompt Engineering.
Next: Retrieval-grounded prompting
✓ Mark this node as complete
Prerequisites Completion of the previous stage and a working local practice environment.
Learning objectives Explain the stage-specific model and its limits Apply the listed concepts using a controlled exercise Record results, failure cases and a reproducible evidence trail
Topics to learn Query rewriting and context selection Citation format and evidence boundaries Untrusted-data handling Abstention when sources are insufficient Prompt Engineering specialization: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Tools Vector search Document parser Citation checker
Hands-on exercise Test a question-answering prompt against supported and unsupported questions. Context for this route: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Suggested project A grounded QA prompt suite with citation and abstention checks. The outcome must directly demonstrate Prompt Engineering, not merely list tools.
Learning resources Completion criteria The system distinguishes supported information from missing evidence. Evidence must be reproducible and tied to Prompt Engineering.
Next: Evaluation and prompt regression
✓ Mark this node as complete
Prerequisites Completion of the previous stage and a working local practice environment.
Learning objectives Explain the stage-specific model and its limits Apply the listed concepts using a controlled exercise Record results, failure cases and a reproducible evidence trail
Topics to learn Golden datasets and rubric design Exact, semantic and human evaluation Regression and variance analysis Prompt injection and robustness tests Prompt Engineering specialization: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Tools LLM-as-judge with human review Evaluation harness Version control
Hands-on exercise Compare two prompt versions on the same fixed evaluation set. Context for this route: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Suggested project A benchmark report with examples, scores and known limitations. The outcome must directly demonstrate Prompt Engineering, not merely list tools.
Learning resources Completion criteria Improvements are supported by a reproducible test set. Evidence must be reproducible and tied to Prompt Engineering.
Next: Production integration and governance
✓ Mark this node as complete
Prerequisites Completion of the previous stage and a working local practice environment.
Learning objectives Explain the stage-specific model and its limits Apply the listed concepts using a controlled exercise Record results, failure cases and a reproducible evidence trail
Topics to learn Prompt versioning and approvals Token/cost and latency budgets Personal data minimisation Logging without storing sensitive prompt payloads Prompt Engineering specialization: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Tools Secret manager Application logs API gateway
Hands-on exercise Implement prompt configuration with version pinning and redact sensitive fields in telemetry. Context for this route: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Suggested project A production prompt registry and change-control checklist. The outcome must directly demonstrate Prompt Engineering, not merely list tools.
Learning resources Completion criteria Prompts can be rolled back and telemetry excludes secrets and personal data. Evidence must be reproducible and tied to Prompt Engineering.
Next: Capstone: reliable prompt workflow
✓ Mark this node as complete
Prerequisites Completion of the previous stage and a working local practice environment.
Learning objectives Explain the stage-specific model and its limits Apply the listed concepts using a controlled exercise Record results, failure cases and a reproducible evidence trail
Topics to learn Use-case and failure-mode definition Prompt, retrieval and schema integration Security and evaluation report Monitoring and release process Prompt Engineering specialization: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Tools LLM API Evaluation harness Git
Hands-on exercise Run a release-readiness review with normal, ambiguous and adversarial inputs. Context for this route: Track-specific focus: Prompt Engineering. Use the tools and standards named in this path; avoid treating this route as a generic overview of Artificial Intelligence.
Suggested project A documented prompt application with golden tests, rollback and usage guidance. The outcome must directly demonstrate Prompt Engineering, not merely list tools.
Learning resources Completion criteria A reviewer can reproduce results and understand where the workflow must abstain. Evidence must be reproducible and tied to Prompt Engineering.
End of roadmap: Complete the final project and build your portfolio.
✓ Mark this node as complete