Certifications

A guide to certifications and online courses for machine learning, cloud computing, algorithms and SQL.

This page is a guide to certifications and online courses that are useful for people building machine learning and cloud skills. For each one, it describes what is covered and who it suits:

Certificates are not a substitute for projects, but a structured curriculum is an efficient way to close gaps – especially in cloud tooling, which is hard to learn from blog posts alone.

Cloud services

AWS Certified Machine Learning – Specialty
Amazon Web Services

Aimed at practitioners who build ML solutions on AWS. The exam covers four domains: data engineering, exploratory data analysis, modelling, and ML implementation and operations. Expect questions on choosing between built-in SageMaker algorithms and custom containers, tuning hyperparameters, handling imbalanced data and deploying models securely. Suits data scientists with some hands-on AWS experience.
AWS Certified Developer – Associate
Amazon Web Services

Focuses on building and maintaining cloud applications: application life-cycle management, CI/CD pipelines, containers, serverless functions and debugging. Useful for ML engineers who wrap models into APIs and services.
AWS Certified Solutions Architect – Associate
Amazon Web Services

Teaches how to design architectures that are scalable, secure, resilient and cost-efficient, with a broad tour of compute, storage, networking and database services. A good second step for anyone who needs to reason about how an ML system fits into a larger platform.
AWS Certified Cloud Practitioner
Amazon Web Services

The foundational certificate: core services, pricing, shared responsibility and basic security concepts. It requires no technical background and works well as a first contact with cloud computing.

Machine learning

Machine Learning Engineer Nanodegree
Udacity

Bridges the gap between notebooks and production. Topics include software engineering practices, object-oriented Python, packaging code, and deploying models with Amazon SageMaker, API Gateway and Lambda. The programme ends with a capstone project – the retinopathy detection post is a good example of the scope such a project can have.
Deep Learning Nanodegree
Udacity

Covers the theory behind neural networks and has learners implement architectures from scratch before moving to PyTorch: CNNs for images, RNNs and LSTMs for sequences, and GANs for generation. Suits people who know classical ML and want a guided path into deep learning.

Coding

Algorithmic Toolbox
University of California San Diego, on Coursera

Key algorithmic techniques that appear again and again in practice: sorting and searching, divide and conquer, greedy algorithms, dynamic programming and recursion. Assignments are graded automatically on time and memory limits, which trains the habit of thinking about complexity.
Data Structures
University of California San Diego, on Coursera

Arrays, linked lists, stacks, queues, hash tables and trees, with an emphasis on when to use which structure and what each operation costs. A natural follow-up to the Algorithmic Toolbox.
Data Structures and Algorithms with Python
Codecademy

A lighter, practice-oriented review of the same material in Python, with interactive exercises. A good refresher before technical interviews.

Databases

SQL for Data Science
University of California Davis, on Coursera

SQL from the data scientist's perspective: selecting, filtering, sorting and aggregating data, then subqueries and joins across tables. Most real projects start with a query rather than a CSV file, so this is time well spent for anyone who learned data analysis in pandas first.