Driven by the mathematical foundations of deep learning, passionate about neural network research and computer vision, and motivated by solving challenging research and engineering problems through building robust, high-quality AI systems.

LinkedIn · Email · Website

I am an aspiring Deep Learning Engineer with a strong interest in the mathematical foundations of machine learning, neural network architectures, and computer vision.

My interests lie in understanding how deep learning models work at a fundamental level, exploring their theoretical foundations, and translating research ideas into practical, well-engineered implementations.

I enjoy studying research papers, investigating challenging problems in neural networks, and developing robust, modular, and reproducible machine learning systems.

I am particularly interested in bridging the gap between theoretical understanding, experimental research, and practical AI engineering.

My work and interests revolve around:

I value:

• Mathematical Rigor: Understanding the theoretical foundations behind machine learning algorithms.
• Research-Driven Development: Making informed engineering decisions through scientific papers, technical documentation, and thorough investigation.
• Experimental Discipline: Systematically configuring, tuning, and evaluating models through controlled experiments.
• Clean Architecture: Building modular, maintainable, and well-documented software systems.
• Engineering Excellence: Solving complex implementation, optimization, and deployment challenges.
• Reproducibility: Ensuring experiments, configurations, and results are traceable and repeatable.

I value mathematical rigor, research-driven development, and systematic experimentation when building machine learning systems.

Python · C# · JavaScript

Pytorch · JTX · Torchvision · Torchinfo

Git · Linox · Docker · CUDA . Github

W&B · Matplotlib · ArXiv . Jupyter Notebook . Colab

My primary areas of interest include:

• Deep Learning Foundations — Exploring the mathematical principles and theoretical foundations underlying deep learning models.
• Neural Network Architectures — Studying neural network structures, learning mechanisms, and architectural innovations.
• Computer Vision — Investigating visual representation learning, image understanding, and modern vision architectures.
• Deep Learning Research — Exploring research papers, evaluating novel methods, and investigating challenging problems in machine learning.

An optimized handwritten digit recognition system built around a custom neural network architecture. The project focuses on classification accuracy, computational efficiency, fast inference, and systematic experimental optimization.

• Approach: Custom Neural Network Architecture, Model Optimization, Hyperparameter Tuning
• Technologies: Python, PyTorch, Weights & Biases (W&B)
• Repository: MNISTVision

A mathematical learning resource focused on the foundations of deep learning. It explores essential mathematical concepts, including linear algebra, differential calculus, optimization, statistics, and probability, with practical explanations and PyTorch implementations.

• Approach: Mathematical Foundations of Deep Learning, Conceptual Analysis, Practical Implementation
• Technologies: Python, PyTorch, NumPy, Matplotlib
• Repository: Axiom

A from-scratch implementation of a handwritten digit recognition model using NumPy. The project focuses on understanding the underlying computations of neural networks through implementation without relying on deep learning frameworks.

• Approach: Neural Network Implementation from Scratch, Convolutional Neural Networks (CNNs), Numerical Computation
• Technologies: Python, NumPy
• Repository: MNISVision-Numpy

I believe in building machine learning systems that are:

• Reproducible: Experiments should be traceable and repeatable.
• Modular: Components should be maintainable and extensible.
• Reliable: Implementations should be tested and evaluated carefully.
• Transparent: Experimental results should be reported honestly.
• Scalable: Systems should be designed with future requirements in mind.

• Working on: MNISVision-Numpy — Implementing and improving a handwritten digit recognition model using NumPy.
• Exploring: Deep Learning fundamentals, neural network architectures, and mathematical foundations.
• Learning: Deep Learning concepts through studying research papers and technical documentation.
• Building: A deeper understanding of neural networks through hands-on implementation, experimentation, and research.

I'm interested in collaborating on:

• Deep Learning Projects: Designing, developing, and implementing neural network models.
• Website Development: Building modern, responsive, and well-structured websites.
• AI-Powered Applications: Developing intelligent applications and integrating machine learning into real-world solutions.

For professional inquiries and technical discussions:

aliasghare1.ai@gmail.com

_{Research · Engineering · Experimentation}