Eight assignments in modern C++ and Python: numerical methods, linear algebra, a neural network from scratch, custom data structures, object-oriented design, PCA face recognition and functional Python.
Course: Advanced Programming (Dr. Amir Jahanshahi), Amirkabir University of Technology. The C++ assignments are graded with GoogleTest unit tests, and every assignment ships with its own Dockerfile.
| # | Folder | Topic | Highlights |
|---|---|---|---|
| 1 | 01-gradient-descent |
Gradient descent | Trains a linear model that predicts student grades; saves and loads weights. |
| 2 | 02-matrix-linear-algebra |
Linear algebra | Matrix class with determinant, inverse and transpose, CSV I/O, and matrix-chain multiplication with dynamic programming. |
| 3 | 03-neural-network |
Neural networks | Feed-forward network built from scratch; compares architectures with different hidden-layer sizes. |
| 4 | 04-circular-linked-list |
Data structures | Priority doubly circular linked list of student records, with search and in-place updates. |
| 5 | 05-midterm-queue-simulation |
Templates and inheritance (midterm) | Generic APQueue; students with normally or exponentially distributed grades, and probabilities by numerical integration. |
| 6 | 06-storehouse-oop |
Object-oriented design | Storehouse management system with a goods hierarchy (food, materials, appliances) and customers. The python/ part covers byte-level image encoding with map and an iterator class. |
| 7 | 07-eigenfaces-pca |
Machine learning in Python | Eigenfaces face recognition with PCA and k-NN on the AT&T faces dataset. |
| 8 | 08-python-functional-decorators |
Functional Python | One-liners with map, filter and reduce; timeit, noexcept and static_vars decorators; a PyQt5 GUI with live plots on a background thread. |
Both course projects are in their own repositories:
- Midterm: 2×2 Rubik's Cube Solver: graph search with BFS, DLS, IDS and bidirectional search.
- Final: Course-Specific Media Player: a PyQt5 lecture player with clickable topic tags.
The C++ assignments share one layout:
0X-topic/
├── cpp/ # implementation, main.cpp and *_unittest.cpp
├── h/ # headers
├── staff/ # figures used in the assignment statement (stuff/ in some folders)
├── README.md # assignment statement
├── Report.pdf # my write-up
├── Makefile
└── Dockerfile # builds GoogleTest and runs the unit tests
Each Dockerfile installs GoogleTest, builds the code and runs the tests:
cd 02-matrix-linear-algebra
docker build -t ap-hw2 .
docker run --rm ap-hw2To build natively instead, install GoogleTest (/usr/local/lib/libgtest.a) and run make in the assignment folder.
The Python assignments (7–8) list their dependencies in the folder (requirements.txt or the Dockerfile).
Mohammad Javad Amin · GitHub