CS 690: AI-Assisted Software Engineering

A project-based course for engineers and technical leaders learning to specify, orchestrate, verify, and ship software with modern AI coding assistants and agentic systems.

Ying Wu College of Computing

Emerging Technologies | Jersey City | Fall 2026

Apply now View syllabus Explore topics

Instructor Mohit Dale

Location Jersey City

Outside effort 8-12 hrs/week

Format Class Meetings + Project

Course focus

Build the judgment layer above the tools.

Students work across the full software lifecycle while learning where generative AI accelerates engineering work, where it creates risk, and how to prove the difference.

01 Specify Requirements, intent, contracts
02 Orchestrate Assistants, tools, agents
03 Verify Tests, types, static analysis
04 Ship CI/CD, cloud, provenance

From prompt to precise intent

Write specs that AI systems can execute against.

Practice requirements elicitation, user stories, acceptance criteria, type signatures, invariants, and formal contracts that reduce ambiguity before code generation begins.

Human-in-the-loop agent design

Coordinate AI coding assistants like an engineering lead.

Explore ReAct, Reflexion, tool use, memory, repository context, and multi-agent workflows using production model APIs and emerging software engineering frameworks.

Evidence before confidence

Build safety nets for AI-generated code.

Combine type checking, unit tests, property-based tests, fuzzing, static analysis, mutation scores, and targeted formal reasoning into a layered validation pipeline.

Professional delivery

Release a working system with traceable AI contributions.

Teams deliver a sophisticated software system with CI gates, security review, deployment, technical reporting, live demonstration, and a provenance ledger for AI-assisted work.

Semester project

The project is the spine.

Teams of three or four build, review, and maintain an industrial-scale software system primarily through AI-assisted development.

Week 3

Proposal

Scope, success criteria, and verification strategy approved before implementation begins.

Week 7

Milestone 1

Architecture review, draft specifications, agent harness, prompts, and test skeleton.

Week 12

Milestone 2

End-to-end feature slice, CI with static analysis, tests, security gates, and risk register.

Week 15

Final release

Tagged release, technical report, live demo, deployment, and AI provenance ledger.

Tooling

Hands-on with the modern AI engineering stack.

Students use agent-capable coding assistants, programmatic access to frontier LLM APIs, GitHub Classroom, Python 3.11 or newer, Docker, static analysis, typed toolchains, and testing frameworks.

Claude Code / Cursor / Aider or equivalent Anthropic, OpenAI, or comparable LLM APIs GitHub Classroom and pull-request workflows pytest, Hypothesis, mypy or pyright, ruff Docker, CI/CD, cloud deployment patterns Threat models, security review, provenance logs

Learn more