The whole degree,
in plain English.

You know Python. C is the one that teaches you what Python was hiding: pointers, manual memory, the stack vs the heap, what a "string" really is. Six weeks of C makes every other language legible. Don't skip it for comfort.

The grammar of computer science: sets, logic, proofs, counting, graphs, recursion. This is the actual prerequisite for Algorithms, and the reason people bounce off Algorithms is almost always this gap, not the algorithms themselves.

Arrays, linked lists, stacks, queues, trees, graphs, hash tables — plus the cost of each operation. The single highest-return course in the entire degree. Implement every one of them from scratch in C, then never again.

Classes, inheritance, polymorphism, interfaces. Less about C++ syntax and more about where to put things so a 40-file codebase doesn't collapse. This is the vocabulary every team you hire will already speak.

Divide-and-conquer, greedy, dynamic programming, graph search, and Big-O analysis. The course that separates "it works on my 10 rows" from "it works on 10 million." Also the entire content of most technical interviews.

Relational modelling, normalization, SQL, transactions, indexing. Nearly every product you'll build is a database with a UI stapled on. Learn normalization properly once and you'll avoid the data mess that kills v2 of most apps.

Named solutions to problems you'll otherwise re-solve badly: Strategy, Observer, Factory, Adapter, Facade, State. Short course, enormous payoff. Officially final-year; there's no reason to wait.

Processes, threads, scheduling, deadlock, memory paging, file systems. Why your server hangs, why two threads corrupt the same variable, what "out of memory" really means. Deploying anything without this is guesswork.

TCP/IP, the layer model, IP addressing and subnets, DNS, HTTP, routing, congestion control. Every "the API is down" incident is somewhere in this course. Skip the older LAN material (token ring, FDDI) — it's history.

Requirements, lifecycle models, UML, testing strategy, configuration management, team process. The syllabus is dated in places (heavy waterfall), but the modelling half — use cases, class and state diagrams — is exactly how you spec work for other people.

Unit testing frameworks, refactoring, build tools, iterative delivery. Tiny credit weight, but this is the day-to-day operating manual for shipping. Pair it with 4513 and treat them as one course.

Vectors, matrices, rank, eigenvectors, SVD, projections. Machine learning is linear algebra with a loss function attached. Strang's book (already the listed text) is the single best resource in this entire syllabus.

Distributions, expectation, variance, Bayes, estimation, hypothesis testing. Needed for ML, for A/B tests, and for not being fooled by your own dashboards. The second half — confidence intervals, significance — is the business-critical part.

Supervised and unsupervised learning, regression, classification, clustering, neural networks, reinforcement learning. Broad, and the syllabus's neural-net section is pre-transformer — pair it with current material, but the fundamentals here are permanent.

Search, heuristics, knowledge representation, planning, game playing. The classical-AI half of the field. Russell & Norvig is the text and it's genuinely worth owning. Ignore the LISP/PROLOG framing — the ideas transfer.

HTML/CSS/JS, HTTP, forms, sessions and cookies, server-side rendering, database tiers. The syllabus's stack is dated (PHP, Flash, AJAX-era) — take the architecture, use a current toolchain for the code.

Virtualization, IaaS/PaaS/SaaS, scaling, distributed storage, map-reduce. This is where your app actually lives and where your monthly bill comes from. Directly commercial knowledge.

Symmetric and public-key crypto, hashing, certificates, TLS, authentication, firewalls. You don't need to invent crypto — you need to stop misusing it. Covers exactly the mistakes that leak customer data.

The graduate-level sequel to Algorithms: randomized algorithms, amortized analysis, network flow, NP-completeness, approximation. Take it when you want depth. Optional for shipping, essential for research.

Structure, LaTeX, figures, citation, plagiarism. Undervalued at 0.75 credits. The ability to write a clear spec, a readable proposal, or a publishable paper compounds harder than almost any technical skill on this list.