Skip to content

Generative AI

Generative AI tools (such as ChatGPT, Copilot, Claude, Gemini, and image-generation models) are now part of professional engineering work. Engineers use AI to brainstorm ideas, explore design options, write and review code, generate documentation, and accelerate analysis tasks.

But there’s a catch: you are still responsible for the correctness, safety, legality, and professionalism of anything the AI helps you create.

Generative AI refers to systems that create new content—text, images, code, diagrams—based on statistical patterns learned from large datasets. For example, ChatGPT can produce code, explain APIs, write documentation, or outline design approaches.

  • Brainstorming ideas and exploring alternatives
  • Drafting code, tests, documentation, emails, and plans
  • Explaining unfamiliar topics or code snippets
  • Generating sample data, messages, or interaction flows
  • Creating diagrams, wireframes, and content outlines
  • Accelerating “first drafts” of writing or implementation
  • Exercise engineering judgment
  • Apply local constraints (e.g., OSU environment, your team’s architecture, your client’s requirements)
  • Produce secure, efficient, or trustworthy code without oversight
  • Make ethical decisions
  • Guarantee correctness: AI frequently hallucinates information
  • Understand the nuances of your Capstone’s unique context
  • Incorrect or misleading information
  • Fabricated citations or nonexistent libraries
  • Insecure or poorly optimized code
  • Biased or inappropriate content
  • Privacy and intellectual property issues
  • Over-reliance leading to lack of understanding

Things change rapidly in the AI world. New models are released frequently, each with its own strengths and weaknesses.

Consider looking at the latest benchmarks and comparisons to choose the right model for your needs.

Some resources to get you started:

Here are some relevant metrics to consider:

As students, you can access GitHub Education for free. It includes access to Copilot Pro and other perks.

Here’s a summary of popular tools and their best use cases:

ToolBest forNotes
ChatGPT / Claude / Gemini / GrokGeneral reasoning, brainstorming, code, documentationGreat for iterative prompting and deep reasoning
GitHub Copilot / Cursor / Claude Code / ChatGPT CodexCode completion, refactoring, tests and moreWorks best when installed in your IDE or Terminal
Nano-Banana / Midjourney / Stable Diffusion / DALL·EImages, diagrams, UI mockupsCheck licenses for generated assets
AI search tools (Perplexity, Bing Copilot)Research, summariesVerify all facts

Most of the recommendations below are supported by most code editors and AI agents. Check your tool’s documentation for specifics.

Tooling is evolving rapidly. New features and capabilities are being added frequently. Stay updated with the latest developments in AI-assisted coding tools to leverage new functionalities as they become available.

This page might not be updated frequently.

Some zero-config tools do exist. For example, for web projects, consider Ultracite.

Setup custom instructions or prompts for your AI code editor.

These instructions will guide the AI in generating code that fits your project’s needs and style.

When providing instructions, consider including:

  • Project Overview: A brief description of your project, its goals, and its target audience.
  • Technology Stack: Specify the programming languages, frameworks, and libraries you are using.
  • Coding Standards: Outline any specific coding conventions or best practices you want the AI to follow.
  • Constraints: Mention any limitations or constraints, such as performance requirements or compatibility issues.
  • Testing Requirements: Specify if you want the AI to generate tests along with the code.
  • Documentation: Indicate if you want the AI to include comments or documentation in the generated code.
  • Collaboration Style: Describe how you want the AI to interact with you (e.g., ask questions, provide explanations, etc.).
  • Persona: Define a persona for the AI to adopt, such as a senior developer, a code reviewer, or a mentor.

Setup memory for your AI code editor.

Memory allows the AI to retain context over time, making interactions more coherent and relevant to your project’s history and goals. Ask your AI code editor or agent to remember important details about your project, preferences, or specific requirements in a separate reference file.

Setup MCP servers to empower your AI code editors.

Some examples:

  • context7 provides up-to-date documentation to your agent
  • Tavily allows your agent to browse the web

Many other options exist. Explore and choose the ones that best fit your needs. Tell your models how and when to use them.