AutoSE

Course plan

Seven key areas, class activities, assignments, and a team project.

  1. 01

    Foundations of Agentic AI

    AI agents, reasoning and planning, memory, tool use and MCP, AI-native software engineering workflows, open-source models, and frameworks.

  2. 02

    Agentic Requirements Engineering

    Specification-driven development: eliciting intent, turning it into clear and testable specifications, and using agents to preserve traceability through implementation.

  3. 03

    AI-assisted Software Architecture

    Architecture knowledge management, ADRs, MCP for architecture tasks, views and viewpoints, architecture evaluation, and architecture copilots.

  4. 04

    Code Generation and Maintenance

    The anatomy of a coding agent, repository-aware generation, agent efficiency, green-field and brown-field work, refactoring, and legacy systems.

  5. 05

    Evaluation of AI-assisted SE

    Artifact evaluation, trustworthiness, sustainability, workflow benchmarks, maintainability, human oversight, governance, and human–automation trade-offs.

  6. 06

    Agentic Testing and Autonomous CloudOps

    Test generation, testing generated artifacts, issue resolution agents, deployment pipelines, agentic root-cause analysis, and autonomous CloudOps.

  7. 07

    Emerging Directions

    Runtime self-adaptation, adaptive orchestration, AI-native IDEs, self-coding information systems, and future directions in AI-assisted software engineering.

Course components

  • Class activities
  • Take-home assignments
  • Team project

Course outcomes

  1. CO1Understand the principles of Agentic AI and their application to software engineering workflows.
  2. CO2Design AI-assisted workflows and agent-centric solutions for software engineering activities.
  3. CO3Apply open-source models, agentic frameworks, and orchestration strategies to automate software engineering and operations tasks.
  4. CO4Critically evaluate AI-generated software artifacts for reliability, performance, trustworthiness, and sustainability.
  5. CO5Deploy and assess AI-assisted software engineering solutions across development and operational workflows.