As data plays a larger part in every organisation, especially digital ones, it becomes more and more necessary to understand where data comes from, how it's stored, and how it's manipulated.

Module 1 - Data collection.
    Introduction to data pipelines
    Introduction to APIs
    HTTP requests
    Selenium for web scraping
    Performing basic web automation actions
    Referencing HTML elements using Xpath
    Cron

Module 2 - Data formats & the Pandas Python library.
    CSV, JSON, Parquet file formats
    Intro to Pandas
‍
Module 3 - Data cleaning.
    Reasons for and approaches to data cleaning
    Handling missing data
‍
Module 4 - Intro to Cloud.
    Data Lakes and Warehouses
    AWS S3
    AWS DynamoDB
    AWS RDS
‍
Module 5 - SQL.
    SQL basics
    SQL Join operations
    SQL Aggregations
    SQL Subqueries
‍
Module 6 - ETL, Distributed Computing & Data Versioning
    Intro to DataBricks
    Distributed computing with PySpark
    ETL
    Data Versioning

This chapter is a comprehensive introduction to data science. We'll start by learning how to explore your data and visualise it, through introducing the industry stardard tools for doing so.

Modelling data is a key part of data science, and a key tool for doing this is machine learning (ML). There will be a deep focus on being practical - actually building every algorithm you learn and knowing how and when to apply it. We’ll begin with an introduction to the basic machine learning problems and the different types of ML. In successive sessions, you'll learn the theory behind a particular algorithm before implementing it and using it on a real dataset.

In this unit, we'll also introduce deep learning - a class of models for representing more complex relationships withing data, especially unstructured data like images.
‍
Module 1 - Exploratory data analysis
    Descriptive statistics
    Data visualisation  

Module 2 - Introduction to machine learning
    What is ML?
    When (NOT) to use ML?
    Linear regression
    Scikit-learn
    Validation and testing
    Hyperparameters, grid search and K-fold cross validation
‍
Module 3 - Theory
    Evaluation metrics
    Fitting Polynomials and the curse of dimensionality
    Bias & variance, underfitting and overfittting
    Regularisation
    Maximum Likelihood Estimation (MLE)
‍
Module 4 - Supervised models
    Classification
    Multiclass classification
    Decision trees

Module 5 - Ensembles
    Random forests

Module 6 - Unsupervised learning techniques
    K-means clustering
    PCA and t-SNE

Module 7 - PyTorch
    Automatic differentiation
    PyTorch Datasets and DataLoaders
    PyTorch Lightning
    Making custom datasets

Module 8 - Deep learning
    Neural networks
    Dropout
    Batch Normalisation
    Convolutional Neural Networks (CNNs)
    Optimisation for deep learning

In this chapter, we'll focus on the engineering required to put models into the real world.

‍Module 1 - Acceleration
    GPUs for PyTorch
    AWS EC2
    Cloud training
‍
Module 2 - Running experiments
    AWS Sagemaker for training
    Ray Tune
‍
Module 3 - Pretrained models
    Pretrained models
    Transfer learning
‍
Module 4 - Building cloud solutions
    AWS Sagemaker for deployment
    AWS Lambda
    AWS API Gateway
    Demo building an end-to-end cloud solution

As the AI systems of more companies mature, there becomes a need for an emerging role that crosses ML engineering with DevOps - ML Operations Engineer, or MLOps Engineer for short. This role is responsible for building, deploying and maintaining ML infrastructure that other teams work with.
‍
Module 1 - Deployment
   Testing using PyTest
    Building an API
    Docker
    TorchServe
‍
Module 2 - Automated deployment, CI/CD and scaling
    Kubernetes
    Kubeflow
    AWS CloudFormation
    GitHub Actions for CI/Cd
‍
Module 3 - Monitoring & continuous training (CT)
    Prometheus
    Grafana
    Thanos
   AWS Sagemaker for monitoring
    Continuous training