This module shows how professional developers use AI coding assistants to work safely with unfamiliar and legacy code. Centered on the agentic loop — a repeatable six-step method (analyze, prompt, review, refine, validate, document) — it teaches the judgment that turns fluent AI output into reviewable, auditable, production-grade work. Learners practice prompt engineering for developers, critical code review of AI-generated code, and disciplined codebase navigation across Java, Python, and mixed-language systems, and learn when a fast answer is enough and when a change demands the full loop. The module builds a shared vocabulary for AI-assisted development, agentic workflows, and human-in-the-loop accountability — relevant to developers, QA engineers, architects, and team leads. IBM Bob serves as the working demonstration tool.

This module teaches developers and the people who guide them how to read unfamiliar code with intent and improve its structure safely. It centers on code refactoring as professional engineering judgment: diagnosing why a function resists change, naming the problem in the shared vocabulary that developers, tech leads, and QA all use, and applying the SOLID principles — the Single Responsibility Principle first — to reduce technical debt without altering behavior. Learners see how the six-step agentic loop turns an AI coding assistant from a code generator into a disciplined refactoring partner, and they practice the critical code review that keeps AI-generated changes safe to merge. The skills transfer across Java, Python, and mixed-language codebases, and they support legacy code modernization at enterprise scale. Whether you write the code or direct those who do, you leave able to judge when a refactor is worth doing and whether it was done well. IBM Bob serves as the working demonstration tool.

This module teaches developers, tech leads, architects, and compliance officers how to modernize legacy Java safely, using a disciplined five-stage modernization lifecycle: assess, plan, execute, validate, and document. It treats legacy code modernization as engineering judgment rather than mechanical translation — assessing what a system guarantees before changing it, planning a Java migration as small releasable slices, validating each slice at an acceptance gate, and recording intent in architecture decision records (ADRs). Learners see how an agentic AI workflow accelerates a version upgrade while a human governs risk, and how to tell architectural awareness from syntactic translation. The skills transfer across Java, Python, .NET, COBOL, and mixed-language codebases, supporting software modernization, code refactoring, and technical debt assessment at enterprise scale. IBM Bob serves as the working demonstration tool.

This module teaches developers and the leads who coordinate them how to scale AI assistance from a single change to an entire codebase. It centers on orchestration: the agentic workflow that splits a system-wide change into stages, delegates each to a specialist role, sequences the rollout across services, and keeps a human at the final checkpoint. Aimed at developers, platform engineers, architects, QA leads, DevOps engineers, engineering managers, IT security, and compliance officers, it builds the judgment that turns AI-assisted development from a pilot into governed, organization-wide use. Learners see how coordinated change, phased rollout, code review, and CI/CD integration keep software modernization consistent and auditable. The skills transfer across Java, Python, JavaScript, .NET, and mixed-language codebases, supporting legacy code modernization, technical debt reduction, and agentic workflows at enterprise scale. Whether you write the code or direct those who do, you leave able to judge when to orchestrate — and when not to. IBM Bob serves as the working demonstration tool.

This module teaches the governance and CI/CD practices that move AI-assisted development from a small pilot to safe, organization-wide use. It centers on governed AI: encoding a team's rules into an AI coding assistant's configuration so they apply the same way everywhere, and keeping a human accountable at every decision that matters. Aimed at developers, architects, QA leads, DevOps engineers, platform engineers, engineering managers, IT security, and compliance officers, it builds the judgment to manage AI adoption across teams without trading speed for risk. Learners see how a governance decision map, a deployment perimeter, CI/CD integration, code review, and an audit trail keep AI-assisted code modernization consistent, secure, and auditable. The practices transfer across Java, Python, JavaScript, .NET, and mixed-language codebases, supporting legacy code modernization, technical debt reduction, agentic workflows, and DevSecOps at enterprise scale. IBM Bob serves as the working demonstration tool.

This is where the whole course comes together. You take one small but realistic legacy codebase — outdated patterns, fragile areas, thin tests, minimal documentation — through the full five-stage modernization lifecycle: assess, plan, execute, validate, and document. Working in IBM Bob in a no-install Coursera lab, you run a compliance review, scope one defensible change, refactor it behind a human approval gate, and verify nothing broke. You capture every decision in a single living document — plan, change record, validation evidence, and reflection — and self-assess it against a structured rubric. By the end, you will have completed a governed, end-to-end modernization and produced a portfolio-ready artifact you can show an employer or interviewer.