NextScript
Deep Learning

Deep Learning

Build CNNs, LSTMs, Transformers, and generative models — and understand the mathematics, ethics, and real-world implications behind every decision.

0.0
|
Leaders (Ages 18-21) (Age Group)
|
advance
|
0 Sessions
What You'll Learn
Understand gradient descent and backpropagation with genuine mathematical precision
Build deep CNNs with residual connections using the Keras Functional API
Train RNNs and LSTMs for text classification and sequence modeling.
Implement and explain the self-attention mechanism behind every major language AI
Use pre-trained Transformers from HuggingFace for real NLP tasks
Write custom training loops, apply regularisation, and interpret model decisions with Grad-CAM
What You Will Build

Across eight modules you will build a progressive portfolio of deep learning projects: a residual CNN trained on real image data, a bidirectional LSTM sentiment classifier, a from-scratch Transformer encoder, a fine-tuned DistilBERT model, a variational autoencoder, and a working DCGAN. The course culminates in a research-style capstone project — a complete, documented deep learning project on a dataset of your choice, presented and defended to a professional standard. Every notebook is committed to GitHub from session one.

Why This Course
This course is for learners who have already built neural networks and are ready to understand what is actually happening inside them — mathematically, architecturally, and in the real world. Deep learning is not a collection of techniques to apply by rote. It is a set of interconnected ideas about representing information, optimising complex functions, and building systems that generalise. This course approaches those ideas with the rigour they deserve — and connects every concept to real deployments, real research papers, and real ethical questions.
Tools & Technologies
Google Colab with T4 GPU runtime
Python 3.10+, TensorFlow 2.x with Keras — primary framework
NumPy, Matplotlib, Seaborn, Scikit-learn, Pandas — supporting libraries
HuggingFace Transformers & Datasets — pre-trained model access and fine-tuning
Deep Learning
0
credits per session
TOTAL
0 Credits
0 Sessions

Create an account to access full course details and start learning