Start Here: Applied AI & Deep Learning in Action

Format: Instructor-Led
Duration: 4 Days • 32 HoursTBD

Overview

Step into the world of deep learning and applied AI with a hands-on, career-aligned course designed to take you from the fundamentals to real-world application. You’ll start by strengthening your foundation in Python, Git, and machine learning, then dive into neural networks, natural language processing, and AI engineering.

Through practical labs and a capstone project, you’ll implement models in PyTorch, experiment with transformers in Hugging Face, and design ethical AI workflows using LangChain. By the end of the course, you’ll have a portfolio-ready project and the confidence to apply modern AI techniques responsibly and effectively in your work.

Outcomes

Demonstrate core machine learning skills — Build, train, and evaluate supervised learning models using Python and scikit-learn.
Implement deep learning architectures — Use PyTorch to construct, train, and optimize neural networks, including CNNs and RNNs.
Apply natural language processing and LLMs — Represent text data through embeddings, build transformer-based models, and analyze results using Hugging Face pipelines.
Evaluate and improve model performance — Benchmark models using performance metrics, costs, and bias analysis to select the right approach for business use cases.
Engineer responsible AI systems — Apply fairness, transparency, and security principles in model deployment.
Leverage LangChain and vector databases — Build prompt-based chains, create retrieval-augmented generation (RAG) systems, and experiment with context-aware AI applications.
Deliver an end-to-end AI solution — Design, build, and present a complete AI project that integrates technical implementation with ethical and practical considerations.

Prerequisites

There are no formal prerequisites; however, this course is best suited for learners who:

Have prior experience using Python and basic programming concepts.
Are familiar with data analysis or machine learning basics (helpful but not required).
Are comfortable writing simple scripts and ready to deepen technical and applied AI skills.

Ideal for:

Early-career data practitioners
Engineers or developers expanding into AI/ML
Professionals seeking hands-on, job-relevant AI skills

System Requirements

✅ Required Installations

✅ Python 3.9+ installed and working
✅ Jupyter Notebook installed (pip install notebook)
✅ Git installed and GitHub account created
✅ PyTorch installed (pip install torch torchvision torchaudio)
✅ Scikit-learn installed (pip install scikit-learn)
✅ Hugging Face Transformers installed (pip install transformers)
✅ OpenAI API installed (pip install openai) and API key configured
✅ LangChain installed (pip install langchain)
✅ FAISS or Chroma Vector Database installed (pip install faiss-cpu or pip install chromadb)
✅ Zoom Desktop Client installed
✅ Slack installed
✅ Google Chrome browser ready
✅ Canvas access confirmed (login credentials provided after enrollment)

Sanity Check (Run Before Class Starts)

# Check Python and pip
python --version
pip --version

# Verify Jupyter installation
jupyter --version

# Test PyTorch
python -c "import torch; print('PyTorch OK:', torch.__version__)"

# Test Hugging Face Transformers
python -c "import transformers; print('Transformers OK:', transformers.__version__)"

# Test OpenAI API
python -c "import openai; print('OpenAI API OK')"

# Test LangChain
python -c "import langchain; print('LangChain OK')"

Module Repos

Module name	Associated Lab	Tool(s) Used
Foundations in Python and Git for AI	- Introduction to the course experience, expectations, and resources. - Ice breakers - Tech setup if needed	Python, Pandas, Numpy, Jupyter Notebook / Google Colab
Core Concepts of Machine Learning	Baseline Builder: Logistic Regression Train a logistic regression model on an image dataset (e.g., MNIST digits)	Jupyter Notebooks, NumPy arrays and pandas DataFrames
Neural Networks: Concepts and Evolution	From Linear to Neural Replace logistic regression with a simple feedforward neural network using PyTorch.
Deep Learning Toolkits in Practice	Pre-Trained Power-Up Use a Hugging Face or torchvision pre-trained model on the same dataset.	PyTorch, Hugging Face, torchvision
Building Neural Nets	Deeper with PyTorch Implement and train a more robust PyTorch model (add multiple layers, experiment with activation functions, and basic optimization).	PyTorch
Neural Net Architectures	CNNs in Action Apply a CNN to the dataset and compare results with previous models. Optionally, explore transfer learning by fine-tuning a pre-trained CNN.	PyTorch
Model Training, Optimization, and Regularization	Review Session	PyTorch
Foundations of Modern NLP	Bag-of-Words to BERT Represent reviews using CountVectorizer, TF-IDF, and embeddings (Word2Vec, BERT). Run a quick sentiment classification and compare results.
NLP Pipelines	Plug-and-Play Pipelines Use Hugging Face’s pipeline API to run sentiment analysis on reviews. Test different pre-trained models (e.g., DistilBERT vs. RoBERTa).	Hugging Face
Introduction to Transformers	Summarizer in Action Apply a transformer model to summarize a batch of long reviews or translate them into another language.
Evaluating NLP Models	Model Match-Up Compare two transformer models (e.g., BERT vs. DistilBERT) on the same dataset using benchmarks (accuracy, F1, inference speed, cost). Include a short reflection on bias and fairness.
Responsible AI and Model Deployment	Review Session
Fundamentals of Prompting and Chaining	Build Your First Chain Use LangChain to call a chat model, create a custom PromptTemplate, and link prompts together into a simple chain.	LangChain
Vector Databases and QA	Ask the Knowledge Base Load documents into a vector database, run similarity searches, and connect them to a LangChain Q&A pipeline for context-aware answers.	LangChain
Review Session + Capstone Workshop	Applied AI Capstone: End-to-End Solution Workshop Learners will work on their capstone project.
Final Project Presentation	Applied AI Capstone: End-to-End Solution In this capstone, learners will plan, build, evaluate, and present a complete AI solution to a real-world problem of their choice. Final Project Due