Find courses anywhere in the world

MIT OpenCourseWare (Spring 2020, Demaine/Ku/Solomon)

MIT's core algorithms course. It teaches how to model computational problems, design correct and efficient algorithms, and prove asymptotic bounds — organised around data structures, sorting, graph search, shortest paths and dynamic programming.

MIT OpenCourseWare (Spring 2017, Terman/Ward)

Builds a complete digital computer bottom-up: MOS devices and combinational logic, sequential circuits and timing, instruction set architecture, a pipelined 32-bit processor (the Beta), caches, virtual memory and operating-system support.

MIT PDOS (xv6 on RISC-V)

Operating systems taught by reading and extending xv6, a small Unix-like kernel for RISC-V. Every concept — isolation, virtual memory, traps, scheduling, file systems — is studied in real kernel source and then implemented in a lab.

Stanford University (Andrew Ng / Tengyu Ma)

Stanford's graduate-level introduction to machine learning: supervised learning, learning theory, unsupervised learning and reinforcement learning, developed mathematically from probability and linear algebra with programming assignments in Python/NumPy.

Stanford University (Fei-Fei Li, Justin Johnson, Serena Yeung)

End-to-end deep learning for visual recognition: from k-NN and linear classifiers through backpropagation, CNN architectures, training recipes, detection and segmentation, generative models, and self-supervised/transformer vision.

Stanford University (Christopher Manning)

Natural language processing built on deep learning: word vectors, dependency parsing, recurrent and attention models, transformers, pretraining and large language models, alignment, and NLP evaluation.

Stanford University (Nick McKeown, Philip Levis)

The Internet from the application layer down: how packets are switched, how TCP provides reliability and congestion control, how IP routing scales, and how security layers on top — with a lab sequence in which you implement a working TCP in C++.

UC Berkeley (Joe Hellerstein / Alvin Cheung)

How a relational database is built and used: SQL and relational algebra, storage and B+ trees, buffer management, join algorithms, query optimisation, transactions with ARIES recovery, concurrency control, and distributed/parallel query processing.

MIT OpenCourseWare (Gilbert Strang)

Strang's classic treatment of linear algebra through the four fundamental subspaces: elimination, vector spaces, orthogonality and least squares, determinants, eigenvalues, SVD and positive-definite matrices, with applications to graphs, differential equations and data.

MIT OpenCourseWare (David Jerison)

Differentiation and integration of functions of one variable: limits, derivatives and their applications, the fundamental theorem of calculus, techniques of integration, and applications to geometry, physics and series.

MIT OpenCourseWare (Ana Bell)

MIT's first programming course for all majors: computational thinking in Python, control flow, functions and recursion, data structures, testing and debugging, object-oriented programming, algorithmic complexity, and simple search/sort algorithms.

Stanford University (Alex Aiken)

Build a complete compiler for the COOL language: lexical analysis, parsing, semantic analysis and type checking, intermediate code generation, runtime organisation, and optimisation.

MIT PDOS (Robert Morris)

Graduate distributed systems taught from classic and modern papers — MapReduce, GFS, Raft, ZooKeeper, Spanner, FaRM — with four demanding Go labs building MapReduce, a Raft library, a replicated key/value service and a sharded store.

DeepLearning.AI on Coursera (Andrew Ng)

A five-course specialization: neural network foundations, practical improvement (regularisation, optimisation, hyperparameters), structuring ML projects, convolutional networks, and sequence models with attention and transformers. All assignments are graded Jupyter notebooks.

Stanford University (Dan Boneh)

Applied cryptography from the ground up: stream and block ciphers, message integrity, authenticated encryption, key exchange, public-key encryption, digital signatures and protocols such as TLS — with security definitions and reduction proofs.

MIT OpenCourseWare (Michael Sipser)

Sipser's own course on automata, computability and complexity: regular and context-free languages, Turing machines and decidability, reducibility, time and space complexity, NP-completeness, and advanced topics such as probabilistic and interactive proofs.

MIT EECS (Rob Miller, Max Goldman)

How to write software that is safe from bugs, easy to understand and ready for change: static typing, specifications, testing, abstract data types, immutability, equality, concurrency and version control, taught in TypeScript/Java with published readings for every class.

MIT OpenCourseWare (Fall 2016, Dourmashkin)

MIT's first-semester physics course: kinematics, Newton's laws, circular motion, energy, momentum, rigid-body rotation, angular momentum and simple harmonic motion, taught with weekly problem solving and desktop experiments.

MIT OpenCourseWare (Spring 2019, TEAL)

Electrostatics through Maxwell's equations: fields and potentials, Gauss's law, capacitance, DC circuits, magnetic fields, induction, inductance, AC circuits and electromagnetic waves, taught in MIT's studio-physics (TEAL) format.

MIT OpenCourseWare (Fall 2007, Auroux)

Calculus of several variables: vectors and matrices, partial derivatives and optimisation, double and triple integrals, vector fields, line and surface integrals, and the theorems of Green, Stokes and Gauss.

MIT OpenCourseWare (Spring 2010, Mattuck/Miller)

Solution methods and qualitative theory for ordinary differential equations: first-order equations, linear equations with constant coefficients, Fourier series, Laplace transforms, and first-order systems with matrix methods and phase portraits.

MIT OpenCourseWare (Fall 2010, Tsitsiklis)

A rigorous first course in probability for engineers: sample spaces and conditioning, discrete and continuous random variables, expectation and variance, limit theorems, Bayesian and classical inference, Bernoulli/Poisson processes and Markov chains.

MIT OpenCourseWare (Fall 2014, Drennan/Taylor)

Introductory college chemistry from an atomic and biological perspective: atomic structure and quantum principles, periodic trends, bonding and molecular geometry, thermodynamics, chemical and acid–base equilibrium, oxidation–reduction, kinetics and transition-metal chemistry.

MIT OpenCourseWare (Fall 2018, Sive/Martin/Lander)

MIT's core biology requirement: biochemistry and molecular structure, genetics and inheritance, molecular biology of the gene, gene regulation, recombinant DNA and genomics, cell signalling, the cell cycle and cancer, and immunology.

MIT OpenCourseWare (Spring 2007, Agarwal)

The lumped circuit abstraction through analogue and digital electronics: resistive networks, network theorems, nonlinear and MOS devices, small-signal amplifiers, energy storage elements, first- and second-order transients, sinusoidal steady state and op-amps.

MIT OpenCourseWare (Fall 2011, Freeman)

Fundamentals of signals and linear time-invariant systems in continuous and discrete time: convolution, Z and Laplace transforms, feedback and control, Fourier series and transforms, sampling, and modulation.

MIT OpenCourseWare (Fall 2006, Mechanical Engineering)

Statics and deformable-body mechanics for engineers: equilibrium of structures, internal forces, stress and strain, axial loading, torsion, bending of beams, deflection, and material behaviour including elasticity and failure criteria.

Stanford University (Computer Science)

Stanford's second programming course, in C++: abstraction and ADTs, recursion and backtracking, algorithmic efficiency, pointers and dynamic memory, linked structures, trees, graphs and hashing, with weekly programming assignments.

Stanford University (Computer Science)

How systems software works end-to-end: filesystems and inodes, processes and signals, multiprocessing and pipelines, threads and synchronisation, networking and HTTP, and building concurrent servers in C/C++.

MIT OpenCourseWare (Spring 2015, Demaine/Devadas/Lynch)

Advanced algorithm design paradigms: divide and conquer, randomised algorithms, greedy methods and amortised analysis, dynamic programming, network flow, linear programming, intractability and approximation, plus distributed and cryptographic algorithms.

University of Michigan (Charles Severance)

A five-course specialization that takes a complete beginner to working data-handling programmer in Python: language fundamentals, data structures, web APIs and scraping, SQL databases, and a data-visualisation capstone.

University of Michigan (Biostatistics)

A three-course specialization on applied statistics in Python: exploratory data analysis and visualisation, sampling and inference, and fitting regression and multilevel models with statsmodels on real survey data (NHANES).

Stanford University (Spring 2024/2025, Hashimoto & Liang)

Students build a language model end to end: byte-pair tokenizer, Transformer implementation, distributed training, data curation, scaling laws, alignment and efficient inference — every component written from scratch rather than called from a library.

UC Berkeley (Fall 2023, Sergey Levine)

Berkeley's graduate course on decision making with deep networks: imitation learning, policy gradients, Q-learning, actor-critic, model-based RL, exploration, offline RL and their application to robotics and control.

UC Berkeley / FSDL (Sergey Karayev, Josh Tobin, Pieter Abbeel)

A production-ML course: how to take a model from notebook to a monitored, versioned, cost-controlled service — covering project setup, data management, experiment tracking, testing, deployment, monitoring and continual learning.

The Linux Foundation / edX (Chris Pokorni, Neependra Khare)

The Linux Foundation's free Kubernetes course: container orchestration concepts, cluster architecture, installation, the API objects (Pods, Deployments, Services, ConfigMaps, Volumes), and how to deploy and expose a real application.

Google Cloud on Coursera (Google SRE team)

Google's SRE course on defining reliability numerically: service level indicators, objectives and error budgets, how to choose and measure them, and how error-budget policy drives release decisions, alerting and toil reduction.

Northwestern University on Coursera (Kevin M. Lynch)

A six-course specialization built on the free Modern Robotics textbook: rigid-body motions and screw theory, forward and inverse kinematics, velocity kinematics and statics, dynamics, trajectory generation and motion planning, and robot control, with a capstone mobile-manipulation project.

MIT (Russ Tedrake, annual open course notes)

Perception, planning and control for robot manipulation in unstructured environments, taught with the Drake toolbox: geometric pose estimation, deep perception, grasping, differential IK, motion planning, force control and learned manipulation policies.

Stanford University (John Ousterhout / Mendel Rosenblum)

Stanford's full-stack web course: HTML/CSS layout, JavaScript, the DOM and browser model, React components and state, HTTP and REST, Node/Express servers, MongoDB persistence, sessions and authentication, and web security — assembled into a photo-sharing application.

University of Helsinki (Matti Luukkainen, Houston Inc.)

A free, credit-bearing university course on modern JavaScript full-stack development: React, Redux, Node/Express, MongoDB, testing with Jest/Playwright, GraphQL, TypeScript, React Native, CI/CD and containers — assessed entirely by submitted exercises.

Stanford University (Paul Hegarty)

Stanford's mobile development course: the Swift language, SwiftUI's declarative view system, MVVM architecture, animation, gestures, persistence, multithreading and networking, taught by building complete iOS applications lecture by lecture.

Stanford University (Feross Aboukhadijeh)

A course on the security model of the web: the same-origin policy, cookies and sessions, XSS, CSRF, injection, clickjacking, HTTPS and certificates, content security policy, and modern browser defenses — with hands-on exploitation and defense assignments.

Harvard University / edX (Colton Ogden, David J. Malan)

Harvard's game development course: 2D and 3D game principles taught by re-implementing classic games in Lua/LÖVE2D and then Unity — covering sprites, collision detection, tilemaps, state machines, procedural generation, physics, audio, AI and publishing.

UC Berkeley (Ren Ng, annual public course site)

Berkeley's graphics course: rasterization and sampling, transforms and the graphics pipeline, texture mapping, geometry and subdivision surfaces, ray tracing, physically based materials and global illumination, and animation — with four large C++ programming assignments.

UC San Diego on Coursera (Scott Klemmer, Elizabeth Gerber)

An eight-course specialization on human-centered design: needfinding and user research, ideation and storyboarding, low- and high-fidelity prototyping, visual and information design, evaluation with heuristics and usability studies, and running controlled experiments — culminating in a portfolio-grade capstone.

Google on Coursera

A seven-course professional certificate covering the UX process end to end: empathising and defining, wireframing and low-fidelity prototypes in Figma, conducting usability studies, high-fidelity design with design systems, responsive web design in Adobe XD/Figma, and building a three-project portfolio.