Open-source teaching materials for "CS 1.5" (formally CS 5001.5), an experimental studio course that merges CS1 ("Intensive Foundations of Computer Science") with Discrete Structures into a single integrated experience — sitting, as the name suggests, right between CS1 and CS2.
It was piloted in the Align Bridge program at Northeastern University, Vancouver (Fall 2025) — a Master of Science pathway for students coming into computing from non-computing bachelor's degrees.
📄 Read the experience report: "CS 1.5": An Experience Report on Integrating CS1 and Discrete Structures for the AI Era — Ildar Akhmetov, Juancho Buchanan (arXiv:2604.16365)
LLMs can generate correct solutions without a student ever understanding them. CS 1.5 is a response to that reality, built on two principles:
- 🤝 AI Embraced — treat LLMs as collaborators, not adversaries. Shift the emphasis from code generation to code comprehension and system design.
- 🧱 Foundations First — build the underlying logic with paper-based exercises and physical props before delegating implementation to a machine.
The goal is an "intuition for problem-solving that transcends tools" rather than syntax memorization — and the instructor's role shifts from "source of all syntax knowledge" to mentor: modeling judgment, verification, and persistence through messy, real-world problems.
- 4-hour studios that interleave CS1 and discrete-math content in one sitting, with weekly recitations and a flipped-classroom format.
- Sharing circles — short, narrative sessions where students, TAs, and instructors trade stories about their learning that week, building a genuine community of practice.
- Embodied learning — a "dancing bubble sort" to feel out O(n²), matryoshka dolls for recursion, knit-and-tape constructions for graphs.
- Code comprehension over generation — students navigate large (2,000–3,000 line) codebases and use AI to understand systems, not just emit snippets.
- Code Walks as assessment — instead of static rubrics, students explain and defend their work in dynamic, "what-if" sessions that double as a final teachable moment.
Each project bridges a mathematical idea with a real, portfolio-ready application — deliberately not "disposable, terminal-based scripts."
| Project | Math concept | What students build |
|---|---|---|
| Set Theory Lab | Set operations, power sets | A Streamlit app that analyzes hand-drawn Venn diagrams with a multimodal LLM |
| Recursive Flood Fill | Mathematical induction | A 2-player strategy game where costs come from inductive proofs, verified against the implementation |
| D&D Combat Simulator | Probability & the Law of Large Numbers | A Monte Carlo battle simulator over real D&D 5e API data; hand-calculated odds vs. empirical convergence |
The 11-week curriculum keeps ~80% of the original content but reorders it for conceptual synergy — so that "induction proves recursion," math objects become classes, and probability becomes Monte Carlo simulation. Time saved on syntax (it ages well with AI help) is reinvested in logic and system design.
Everything here is intended for educators to adopt, adapt, or learn from. Repositories typically ship starter code only (no solutions), are released under permissive licenses (e.g. CC0 1.0), and several are set up as GitHub templates so you can spin up your own copy or wire them into GitHub Classroom in one click.
If you adopt or remix something here, we'd love to hear about it — open an issue or reach out via the contact details in the paper.
Maintained by Ildar Akhmetov (@ildarakhmetov) and Juancho Buchanan (@sillyfunnypedro).