Skip to contentSkip to main content
Practical AI SkillsBuild real AI skills and apply them to your role with expert feedbackFind My Learning Path
Microcourse Roadmap
By Academy

Microcourses, by Academy

60 of 300 planned courses are drafted so far — each with a specific problem, a clear title, a target role, and a practical output.

IT-001First production shortlistRoadmap — not yet produced

AI Debugging: Find and Fix One Reproducible Bug

Problem: AI-generated code looks right, but the actual cause of a bug isn't clear.

Reproduction steps, a verified fix, and a regression test.

For: Software Developer, Application Support Engineer

Get notified when this launches
IT-002Roadmap — not yet produced

GitHub Copilot: Build One Feature You Can Test

Problem: Part of the requirement gets missed while AI builds a feature.

A small feature checked against acceptance criteria.

For: Software Developer, Full-Stack Developer

Get notified when this launches
IT-003First production shortlistRoadmap — not yet produced

AI Unit Tests: Catch the Missing Edge Cases

Problem: Unit tests cover normal cases, but important edge cases get left out.

Positive, boundary, and failure tests for one function.

For: Developer, QA Automation Engineer

Get notified when this launches
IT-004Roadmap — not yet produced

AI Code Review: Spot Risky Changes Before Merge

Problem: Risky changes or missing checks go unnoticed in pull-request review.

Evidence-backed review comments and verified findings.

For: Developer, Technical Lead

Get notified when this launches
IT-005First production shortlistRoadmap — not yet produced

AI Codebase Navigator: Understand an Existing Module

Problem: It's hard to understand what a given module does in an existing codebase.

A module map and execution flow with file references.

For: Developer, Junior Engineer

Get notified when this launches
IT-006Roadmap — not yet produced

AI Refactoring: Clean One Function Without Breaking Behavior

Problem: Cleaning up messy code breaks existing behaviour.

A refactored function with before/after passing tests.

For: Developer, Technical Lead

Get notified when this launches
IT-007Roadmap — not yet produced

AI API Tests: Check Success and Failure Paths

Problem: API testing doesn't properly check invalid requests and error responses.

A request-test matrix and response assertions for one endpoint.

For: QA Engineer, Backend Developer

Get notified when this launches
IT-008Roadmap — not yet produced

AI SQL Assistant: Write and Verify Read-Only Queries

Problem: AI's SQL query looks plausible but returns incorrect results.

A validated query and expected-result checks on a sample database.

For: Data Analyst, Backend Developer

Get notified when this launches
IT-009Roadmap — not yet produced

AI API Docs: Explain One Endpoint Accurately

Problem: API documentation doesn't match the actual implementation.

Verified request, response, and error documentation.

For: Backend Developer, Technical Writer

Get notified when this launches
IT-010Roadmap — not yet produced

AI Log Triage: Turn Errors into an Investigation Plan

Problem: It's unclear where to start investigating from long error logs.

A sanitized log summary and an ordered investigation checklist.

For: DevOps Engineer, Support Engineer

Get notified when this launches
IT-011Roadmap — not yet produced

AI User Stories: Turn Requests into Testable Tasks

Problem: A vague business request can't be converted into testable development tasks.

A user story, acceptance criteria, and a bounded task breakdown.

For: Business Analyst, Developer

Get notified when this launches
IT-012Roadmap — not yet produced

AI Pull Request Summaries: Explain What Changed

Problem: Reviewers can't tell the purpose and impact of a change from the pull-request description.

A diff-verified summary, risks, and test notes.

For: Developer, Technical Lead

Get notified when this launches
IT-013Roadmap — not yet produced

AI Data Cleanup: Build a Validated CSV Cleaning Script

Problem: Manually cleaning CSV data is repetitive and error-prone.

A cleaning script with before/after data checks.

For: Data Analyst, Python Developer

Get notified when this launches
IT-014Roadmap — not yet produced

AI Output Validation: Turn Model Replies into Checked JSON

Problem: The LLM's response doesn't follow the expected JSON format, and the application fails.

Schema validation and failure handling in a starter endpoint.

For: Backend Developer, AI Developer

Get notified when this launches
IT-015Roadmap — not yet produced

RAG Answer Check: Add Sources and Handle Unknowns

Problem: A document chatbot answers without sources and guesses on unknown questions.

Source references and an unsupported-answer fallback in a starter chatbot.

For: AI Developer, Backend Developer

Get notified when this launches

Focus & delivery boundaries, by academy

IT & Tech

Understanding, debugging, testing, and verifying AI-generated code — not just generating it. In Stack Overflow's 2025 Developer Survey, 66% of respondents found almost-correct AI solutions frustrating, and 45% said debugging AI-generated code was more time-consuming, so this first batch focuses on reliable, everyday engineering tasks.

Scope: IT-014 and IT-015 are starter-template labs, not complete application-development programs — coding/API prerequisites are shown upfront.

AI output is a starting point for review, not an unreviewed commit — every course ends with a verified, tested result.

Reusable kit: Sample repository, prompt pattern, input fixtures, test cases, and a verification checklist.

HR & Talent

Recruitment administration, candidate information organization, interview consistency, and employee operations. LinkedIn's Future of Recruiting report highlights streamlining repetitive recruiting work, skills-based hiring, and the continued importance of recruiter judgment.

Use AI to organize information and prepare drafts — hiring, rejection, compensation, and termination decisions stay human-made, never automatic.

Reusable kit: Synthetic resumes, approved policy samples, editable spreadsheets, interview rubrics, and a human-review checklist.

Marketing

Generic content, inconsistent brand voice, poor repurposing, and unverified reporting. HubSpot's 2026 research highlights challenges with personalized/channel-specific content, brand values, content repurposing, and demonstrating ROI.

Course promise is better briefs, useful content assets, clear offers, and checked analysis — never "guaranteed viral," "guaranteed ranking," or "guaranteed leads."

Reusable kit: Sample brand brief, approved source material, content templates, a creative test sheet, and a publishing checklist.

Sales & CRM

Prospect research, generic outreach, incomplete CRM records, proposal drafting, and unclear follow-ups. Salesforce's 2026 research names sales administration, prospecting bandwidth, coaching gaps, and data quality as key challenges, with 74% of sales professionals reporting a focus on data cleansing.

No live bulk outreach, destructive CRM merging, automatic discount approval, or unconfirmed meeting commitments — every output is a reviewed draft or sandbox workflow.

Reusable kit: Synthetic prospect data, consented/sample transcripts, approved offer documents, CRM templates, and a pre-send checklist.