Academic-Grade AI-Powered Technical Support System
Production-quality code assistance with pedagogical excellence for Computer Science education
Features • Architecture • Quick Start • Usage • Contributing
This is a fun experimental project developed by the author to try vibe coding and built to support CS professors with a $25 yearly annual subscription to build a team of agents using Z.ai GLM-4.7. It is designed to explore AI-powered technical assistance and is not intended to:
- Replace students or their learning experiences
- Serve as a production tool for classroom use
- Provide authoritative answers for coursework
- Replace professional teaching assistants or human tutors
Use this system as a supplementary learning tool and always prioritize student growth over convenience.
The CS Professor Technical Agent Team is a sophisticated multi-agent system designed specifically for computer science professors.
This system provides technical support for:
- Course Development: Curriculum design, assignments, and teaching materials
- Research Support: Implementing algorithms, data analysis, and reproducibility
- Code Quality: Review, testing, and optimization of student and research code
- Pedagogical Value: Clear explanations, learning objectives, and educational examples
- Industry Standards: Following Google Style Guides, Microsoft Best Practices, and Facebook Engineering standards
- Academic Rigor: Meeting publication standards from NeurIPS, ICML, CVPR, and top CS journals
- Reproducibility: Code that can be reproduced and validated
- Performance Optimization: Profiling and optimization for research-scale problems
- Learning Objectives: Aligned with Bloom's Taxonomy and ACM/IEEE curriculum guidelines
- Progressive Complexity: From fundamentals to advanced concepts with proper scaffolding
- Real-World Examples: Connecting theory to industry practice
- Inclusive Design: Universal Design for Learning (UDL) principles
- 16 Specialized Agents: Covering all major CS domains
- Intelligent Coordination: Automatic routing to appropriate experts
- Context Awareness: Understanding academic vs production contexts
- Multi-Disciplinary: Combining algorithms, systems, ML, databases, and more
CS Professor Technical Coordinator (Core Agent)
│
├── Web Development & Architecture
│ ├── Web Developer (Full-stack, APIs, Modern Frameworks)
│ ├── Database Expert (SQL/NoSQL, Performance, Design)
│ └── DevOps Engineer (CI/CD, Infrastructure, Deployment)
│
├── Algorithms & Systems
│ ├── Algorithm Specialist (Design, Analysis, Optimization)
│ ├── Systems Programmer (OS, Low-level, Performance)
│ └── Performance Profiler (Optimization, Profiling, Benchmarks)
│
├── Machine Learning & AI
│ ├── AI/ML Developer (Deep Learning, Training, Deployment)
│ └── Statistician (Data Science, Analysis, Experimental Design)
│
├── Code Quality & Engineering
│ ├── Code Reviewer (Best Practices, Standards, Quality)
│ ├── Research Code Reviewer (Reproducibility, Publication Standards)
│ ├── Debugging Specialist (Troubleshooting, Analysis, Fixing)
│ └── Testing/QA Engineer (Test Strategy, Automation, Coverage)
│
├── Educational & Academic
│ ├── Education Specialist (Pedagogy, Curriculum, Teaching)
│ ├── Testing & Grading (Automated Assessment, Feedback)
│ └── Git & Version Control (Collaboration, Best Practices)
│
└── Security & Infrastructure
└── Cybersecurity Specialist (Secure Coding, Threat Analysis)
- Node.js 18+ or Bun 1.0+
- Git (for version control examples)
# Clone the repository
git clone https://github.com/yourusername/cs-prof-agents.git
cd cs-prof-agents
# Install dependencies
bun install
# or
npm install
# Agents are ready to useSimply describe your technical task, and the coordinator will delegate to appropriate experts:
"Help me create an assignment on graph algorithms for a junior-level algorithms course.
Requirements:
- Implement Dijkstra's algorithm and A* search
- Compare performance on random graphs
- Include test cases and grading rubric
- Provide student-friendly explanations
- Difficulty should be appropriate for CS students"
"I need to implement a transformer model for natural language processing.
Requirements:
- Follow the 'Attention Is All You Need' architecture
- Implement from scratch in PyTorch
- Include proper documentation for reproducibility
- Performance-optimized for training on large datasets
- Ready for publication-quality code"
"Review this student's implementation of a binary search tree.
Focus on:
- Correctness of insert, delete, and search operations
- Code quality and style (following Google Python Style Guide)
- Common pitfalls and edge cases
- Pedagogical suggestions for improvement
- Constructive feedback suitable for learning"
"This Python script for processing large CSV files is taking too long.
It processes 10M rows in 30 minutes.
Help me:
- Profile the code to identify bottlenecks
- Optimize for better performance
- Provide before/after performance comparison
- Explain the optimizations for students
- Maintain code readability"
You can also directly invoke specific subagents for specialized tasks:
# Get web development help
@.agents/agent/subagents/web-developer.md
# Request code review
@.agents/agent/subagents/code-reviewer.md
# Database optimization
@.agents/agent/subagents/database-expert.md
# Curriculum design
@.agents/agent/subagents/education-specialist.md
# Research code review
@.agents/agent/subagents/research-code-reviewer.mdRole: Central coordinator for all technical tasks with academic awareness
Expertise:
- Technical architecture and design decisions
- Technology selection and best practices
- Multi-disciplinary project coordination
- Quality standards and pedagogical integration
- Research reproducibility guidelines
Tools: write, edit, bash
Temperature: 0.2 (analytical and precise)
| Agent | Specialization | Example Tasks |
|---|---|---|
| Web Developer | Full-stack development, APIs, modern frameworks | RESTful APIs, React/Vue components, authentication systems |
| Systems Programmer | Low-level systems, OS internals, performance optimization | Thread pools, memory allocators, concurrent data structures |
| Database Expert | Database design, SQL/NoSQL optimization, query performance | Schema design, indexing strategies, query optimization |
| DevOps Engineer | CI/CD pipelines, infrastructure as code, deployment | GitHub Actions workflows, Kubernetes deployments, Terraform modules |
| Agent | Specialization | Example Tasks |
|---|---|---|
| Algorithm Specialist | Algorithm design, complexity analysis, data structures | Graph algorithms, dynamic programming, NP-complete problems |
| Performance Profiler | Performance analysis, profiling, benchmarking | CPU profiling, memory leak detection, optimization strategies |
| Debugging Specialist | Code debugging, error analysis, troubleshooting | Race conditions, segmentation faults, performance bugs |
| Agent | Specialization | Example Tasks |
|---|---|---|
| AI/ML Developer | Machine learning, deep learning, model deployment | Neural networks, training pipelines, model optimization |
| Statistician | Statistical analysis, experimental design, data science | Hypothesis testing, A/B testing, regression analysis |
| Agent | Specialization | Example Tasks |
|---|---|---|
| Code Reviewer | Code quality, best practices, standards | Style guide compliance, design patterns, security reviews |
| Research Code Reviewer | Academic code standards, reproducibility, publication readiness | Reproducibility checks, experimental design, documentation review |
| Testing/QA Engineer | Testing strategies, test automation, coverage | Unit tests, integration tests, test-driven development |
| Agent | Specialization | Example Tasks |
|---|---|---|
| Education Specialist | Pedagogy, curriculum design, teaching strategies | Learning objectives, active learning activities, assessment design |
| Testing & Grading | Automated grading, assessment systems, feedback | Auto-graders, plagiarism detection, rubric design |
| Git & Version Control | Version control, collaboration, best practices | Git workflows, code review processes, repository management |
| Agent | Specialization | Example Tasks |
|---|---|---|
| Cybersecurity Specialist | Secure coding, vulnerability analysis, threat modeling | OWASP compliance, penetration testing, encryption implementation |
Scenario: Creating a new algorithms course
Request: "Design a semester-long algorithms course with weekly assignments"
Response includes:
- Course syllabus with learning objectives
- Weekly schedule with topics and activities
- Assignment designs with progressive difficulty
- Grading rubrics and assessment strategies
- Active learning activities for each topic
- Common student misconceptions and how to address them
- Technology stack recommendations
- Example code solutions with pedagogical annotationsScenario: Implementing a research paper's algorithm
Request: "Implement the ResNet architecture from the Deep Residual Learning paper"
Response includes:
- Complete PyTorch implementation with proper architecture
- Data preprocessing pipeline
- Training loop with proper logging and checkpointing
- Performance optimization techniques
- Reproducibility documentation
- Experimental validation
- Publication-ready code quality
- Comparison with baselinesScenario: Helping a struggling student
Request: "This student is having trouble understanding dynamic programming.
Help me explain it with examples and practice problems"
Response includes:
- Intuitive explanation of DP concepts
- Step-by-step walkthrough of classic DP problems
- Progressive examples from simple to complex
- Common mistakes and how to avoid them
- Practice problems with solutions
- Visualizations and analogies
- Connection to related topicsScenario: Reviewing student submissions
Request: "Review these 10 student submissions for a web API assignment"
Response includes:
- Consistent grading based on rubric
- Detailed feedback for each submission
- Identification of common issues across submissions
- Pedagogical suggestions for addressing common mistakes
- Code quality improvements
- Security considerations
- Performance optimization suggestions- Be Specific: Include requirements, constraints, and context
- Specify Audience: Indicate student level (introductory, intermediate, advanced)
- Provide Context: Share course materials, research papers, or code snippets
- Define Success Criteria: Specify what quality level you expect
- Iterate: Refine requests based on initial outputs
- Combine Agents: Use multiple agents for complex, multi-disciplinary projects
Full-Stack Course Project
- Education Specialist (course context and learning objectives)
- Web Developer (frontend/backend implementation)
- Database Expert (schema design and queries)
- Code Reviewer (quality assessment)
- Testing & Grading (auto-grader and rubric)
- Git & Version Control (repository setup and collaboration)
Research Paper Implementation
- Algorithm Specialist (algorithm understanding and design)
- AI/ML Developer (implementation in ML framework)
- Performance Profiler (optimization and benchmarking)
- Research Code Reviewer (reproducibility and publication readiness)
- Statistician (experimental design and validation)
- Code Reviewer (final quality review)
Student Assignment Grading
- Testing & Grading (auto-grader implementation)
- Code Reviewer (quality feedback)
- Debugging Specialist (identify and explain bugs)
- Education Specialist (pedagogical feedback)
All agents use the GLM-4.7 model for consistent, high-quality responses. Temperature settings are optimized for each agent's purpose:
- Coordinator: 0.2 (precise and analytical)
- Code Reviewer: 0.1 (precision-focused)
- Technical Subagents: 0.2 (accurate and analytical)
- Education Specialist: 0.3 (more conversational and explanatory)
cs-prof-agents/
├── .agents/
│ ├── agent/
│ │ ├── core.md # CS Professor Technical Coordinator
│ │ └── subagents/
│ │ ├── web-developer.md # Full-stack web development
│ │ ├── systems-programmer.md # Low-level systems programming
│ │ ├── database-expert.md # Database design & optimization
│ │ ├── devops-engineer.md # DevOps & infrastructure
│ │ ├── algorithm-specialist.md # Algorithm design & analysis
│ │ ├── performance-profiler.md # Performance optimization
│ │ ├── debugging-specialist.md # Code debugging & troubleshooting
│ │ ├── ai-ml-developer.md # Machine learning & deep learning
│ │ ├── statistician.md # Statistical analysis & data science
│ │ ├── code-reviewer.md # Code quality & best practices
│ │ ├── research-code-reviewer.md # Academic code standards
│ │ ├── testing-qa.md # Testing strategies & automation
│ │ ├── cybersecurity.md # Security & vulnerability analysis
│ │ ├── education-specialist.md # Pedagogy & curriculum design
│ │ ├── testing-grading.md # Automated assessment & grading
│ │ ├── git-version-control.md # Version control & collaboration
│ │ └── network-scientist.md # Network analysis & graph data science
│ ├── examples/ # Usage examples and scenarios
│ │ └── usage-examples.md
│ └── docs/ # Technical documentation
├── package.json
├── bun.lock
└── README.md
We welcome contributions to enhance the CS Professor Technical Agent Team!
- New Specialized Agents: Expert domains not yet covered
- Enhanced Prompts: Improved agent capabilities and responses
- Usage Examples: Additional teaching and research scenarios
- Bug Fixes: Correcting issues and improving reliability
- Performance: Optimization of agent responses and coordination
- Documentation: Improving clarity and completeness
-
Fork the repository
git clone https://github.com/yourusername/cs-prof-agents.git
-
Create a feature branch
git checkout -b feature/your-feature-name
-
Make your changes
- Follow existing agent structure and format
- Include proper frontmatter metadata
- Add examples and use cases
- Update documentation as needed
-
Test thoroughly
bun install
-
Commit and push
git add . git commit -m "feat: add new specialized agent for domain X" git push origin feature/your-feature-name
-
Submit a Pull Request
- Describe your changes clearly
- Reference related issues
- Include screenshots or examples if applicable
Please be respectful and constructive in all interactions. We are committed to providing a welcoming and inclusive environment.
This project is licensed under the MIT License - see the LICENSE file for details.
- Issues: Report bugs and feature requests on GitHub Issues
- Discussions: Join community discussions on GitHub Discussions
- Email: For inquiries, contact [maintainer@example.com]
This system is inspired by:
- Google Engineering practices and documentation standards
- Microsoft Research and their contribution to CS education
- Meta AI Research and their open research practices
- Top CS programs at MIT, Stanford, CMU, UC Berkeley
- Open source community and their collaborative spirit
If you find this system helpful for your teaching or research, please consider giving it a ⭐ on GitHub!
Built with ❤️ for CS professors