The Artificial Intelligence (AI) Developer’s Blueprint: From Modeling to PyTorch
Continuing Education Certificate Course
May 2–3, 2026
The College of New Jersey (TCNJ)
Ewing, NJ
Join us for a two-day, in-person certificate course on the Fundamentals of Artificial Intelligence with a clear focus on the mathematics and computational tools that drive modern machine learning. The course, The AI Developer’s Blueprint: From Modeling to PyTorch, is designed for learners who want to understand how AI works at a fundamental technical level instead of relying on packaged software.
Participants will learn how gradients, loss functions, and regularization connect directly to the training of neural networks. Each mathematical idea will be translated into MATLAB or PyTorch code such that participants understand how theory is applied to real-world problems. The goal is fluency: understanding what the algorithm is doing, why it works, and how to adjust it when learning fails. This course is ideal for working professionals, graduate students, and advanced undergraduates who want a strong, practical foundation in artificial intelligence without shortcuts or oversimplifications. Participants leave with skills they can apply immediately and an understanding of AI that will grow with them throughout their careers.
Course Overview
The course will cover many AI topics relevant to neural networks and machine learning. Specifically, it seeks to answer the following questions:
- Introduction to Artificial Intelligence
- What does AI mean in practical engineering terms?
- How is AI different from machine learning?
- How do you build and train a simple AI system from scratch?
- Neural Networks: The Framework for Machine Learning
- How do neural networks process inputs, weights, activation, and outputs?
- What does the perceptron teach about supervised learning and its limits?
- Parametric Modeling
- How does linear regression learn through parameter updates?
- Why do we use logistic regression for classification instead of linear models?
- PyTorch for Machine Learning and the Perceptron
- What does the perceptron teach about supervised learning and its limits?
- How does PyTorch use tensors, AutoGrad, and optimizers for machine learning?
- How do we structure a full machine learning project in PyTorch and control training behavior?
- Image Recognition via Machine Learning
- How can a shallow neural network classify handwritten digits?
- Mathematical Optimization and Gradient Descent
- Why do machine learning models rely on gradient-based optimization?
- When should you use stochastic, mini-batch, or full-batch gradient descent?
- How does backpropagation change each synapse weight during training?
- How do hyperparameters like learning rate, momentum, and weight decay affect convergence?
- Automatic Differentiation: Basis of AutoGrad
- How does automatic differentiation compute gradients efficiently?
- Advanced Topics and Project
Note that the exact topics covered will depend on the needs of participating students and the pace set by the class.
Target Audience
This course is aimed at working professionals, graduate students, and advanced undergraduates who have completed at least two semesters of single variable and integral calculus and one semester of Python-based programming. Individuals with extensive experience in other programming languages may also be well prepared. Additional math, programming, and data science-relevant knowledge will also be helpful.
Why Take This Course
Artificial intelligence influences nearly every major field today, but many courses and textbooks emphasize software usage rather than true understanding. This certificate program is designed for students who want to learn how and why artificial intelligence works, not just how to call a library in Python. The curriculum is driven by teaching priorities rather than a research agenda; every topic was selected because it builds the conceptual foundation that beginners and intermediate learners actually need to progress.
The course does not rush to showcase advanced applications or oversized neural networks. Instead, participants learn AI as a fundamental engineering subject, similar in spirit to signals, circuits, or control. Mathematical theory is introduced with step-by-step derivations, and students carry out key operations manually before moving to automation. Concepts such as gradient descent, loss minimization, and backpropagation are first worked through on paper and in small Excel or MATLAB examples, making the learning process transparent rather than opaque.
Only after the underlying ideas are fully understood do we transition into PyTorch. At that point, students not only know what the code is doing, but why each line exists. The examples remain intentionally manageable so participants can see how individual training samples influence learning, how hyperparameters affect convergence, and how the network evolves over time. This structure prepares students to approach larger models and datasets confidently instead of hoping that a software package performs correctly.
By the end of the program, participants do not simply know how to use artificial intelligence tools. They understand the mathematical and algorithmic mechanisms that govern learning, giving them the ability to evaluate, modify, and improve neural network architectures rather than remain dependent on pre-built solutions. This course is intended for those who want to become creators in AI rather than software operators.
Schedule, Location, and Cost
Four Meetings: 2 In-Person, 2 Virtual
- In-Person Meetings: Saturday 5/2 and Sunday 5/3/2026 9:00 a.m. to 5:00 p.m.
- Virtual Pre-Meeting Session: Saturday 4/25 12:00 p.m. to 2:00 p.m.
- Virtual Post-Meeting Session: Saturday 5/9 12:00 p.m. to 2:00 p.m.
Location: The College of New Jersey Campus, Ewing, NJ
Cost: $2,000 (includes all instruction, materials, and certificate of completion)
** Initial payment of $1,000 due at time of application; Second payment of $1,000 due by April 1st.
Disclaimer – A minimum of 12 students is required to run this course. If the minimum is not met, registrants will be issued a full refund.
Instructor
The course will be taught by Dr. Anthony S. Deese, Professor and Chair of Electrical and Computer Engineering at The College of New Jersey. Dr. Deese is widely known for his YouTube channel with more than 150 instructional videos that make advanced mathematical ideas in AI and machine learning easier to understand for students across the world.
See the following preview of his teaching style:
Questions
If you have any questions, please email the Department of Electrical and Computer Engineering at The College of New Jersey at eceng@tcnj.edu.