If you go online, you will find jobs, courses, and training programs that can move your career forward. But, in what direction? When it comes to choosing the next step in your AI career, there are two things that will help you strategize. First, you have to learn about yourself. Then, you need to explore opportunities that fit for your present skills and will help you build new skills.
By taking the Workera test, you will:
- know where you stand in the community of AI practitioners
- review your performance to learn about your strengths and weaknesses
- access personalized study plan to prepare for interviews
- fast track to job opportunities within our network
- get a certificate
This guide is intended to help you understand what the test evaluates and prepare for it.
I What the test evaluates
The test evaluates your ability to carry out tasks involved in AI projects including data engineering, modeling, deployment, business analysis, and AI infrastructure. The skills required to achieve those tasks are clustered and evaluated in six distinct test sections described in the table below. Each section is comprised of multiple choice questions selected from a large database, so that different test takers get different questions.
|Machine learning||This section evaluates your mastery of ML models (e.g., PCA, K-means, K-NNs, SVM, Logistic Regression, Linear Regression and Decision Trees), methods to train ML models (e.g., initialization, optimization, regularization, and model selection), and decision making in ML projects (e.g., error analysis, collecting and labeling data, splitting and augmenting data, and choosing an evaluation metric).||17 min|
|Deep learning||This section evaluates your understanding of deep learning fouundations (e.g., neurons, fully-connected layers, deep neural networks, forward propagation, backpropagation, and activation functions), techniques to improve neural networks (e.g., initialization, optimization, regularization, and selection of deep learning models), convolutional neural networks (e.g., convolutional and maxpool layers, classic CNNs, and techniques such as skip connections), and sequence models (e.g., recurrent neural networks, LSTM and GRU cells, and methods such as embeddings used in natural language processing, speech recognition or time series analysis).||15 min|
|Data science||This section evaluates your ability to use probabilities (e.g., probability distributions, operations on probabilities, and properties of random variables), use statistics (e.g., statistical theorems like the Central Limit Theorem, statistical parameters, and hypothesis testing), and analyze data (e.g., data preprocessing, data visualization and the use of common metrics).||15 min|
|Mathematics||This section evaluates your ability to work with mathematical objects and operations used in AI. This includes matrix/vector operations and properties (e.g., dot product, outer product, summation, multiplication, transposition, inverse, determinant, etc.), mathematical functions and their properties (e.g. Euclidian norm, L1 norm, sigmoid, ReLU, tanh, softmax, cosine, sine, min, max, argmin, argmax, convexity, concavity, etc.), and understanding calculus (e.g., differentiating and integrating common functions).||14 min|
|Algorithmic coding||This section evaluates your ability to solve algorithmic problems (e.g., sorting, search on trees and graphs, matching, permutation, counting, tree traversal, etc.), use data structures and types (e.g., lists/arrays, sets, maps/dictionaries, trees, graphs, linked lists, tuples, queues, stacks, heaps, strings, integers, floats, etc.), and apply coding methods (e.g., recursion, backtracking, divide and conquer, etc.).||16 min|
|Software engineering||This section evaluates your understanding of software development management methodologies (e.g., version control, management frameworks like Scrum and Agile, and processes to review and test code), design patterns and architectures (e.g., object-oriented design, domain-driven design, monolithic architecture and microservices), computer networking (e.g., TCP-IP, basic encryption, HTTP, Domain Name Server, Cookies, IPv4, IPv8, etc.), databases (e.g., relational database management systems and SQL), and computing (e.g., cloud and distributed computing, performance measures like latency and throughput, concurrency, parallelism, and race conditions).||15 min|
II How the test was built
Our computerized tests are created by a team of experts in AI and assessment led by Dr. Andrew Ng and Kian Katanforoosh from deeplearning.ai and Stanford University. Our massive question library covers more than 500 AI and software topics.
Assessing the right skills
We interviewed more than 100 machine learning and data science leaders to understand what tasks are necessary to develop AI projects. With that information, we derived a list of AI skills required to carry out the identified tasks. Then, a team of experts wrote and continue to write questions to evaluate the relevant skills.
Providing actionable feedback
Questions were grouped in six ubiquitous subjects so that people can identify their strength and weaknesses at the skill and subject levels.
Assuring test quality
The test is monitored both qualitatively and quantitatively:
- Assessment experts ensure that every question undergoes a rigorous QA process before being available for use. Besides, they are attentive to user feedback and periodically review questions in production to improve the test experience.
- Machine learning algorithms monitor the test. They ensure that questions distinguish people of distinct abilities and cover a wide range of difficulties.
III Preparing for the test
Practice makes perfect, so we recommend that you check the following resources to practice for each section of the test. Remember that you have 3 attempts so don’t overstress: use your first attempt to practice.
- Preparing for the machine learning test
- Preparing for the deep learning test
- Preparing for the data science test
- Preparing for the mathematics test
- Preparing for the algorithmic coding test
- Preparing for the software engineering test
IV Tips for taking the test
Here are some tips to ensure a good testing experience:
Find a quiet place, free of distractions, to take the test.
To help you stay focused, don’t leave the test page and don’t have distracting tabs open.
Once you have started a section don’t attempt to close or go back in the browser as you may lose your progress.
You can choose to opt out of the software engineering and deep learning sections. The remaining sections are mandatory in order for us to recommend the most suited AI career pathways and opportunities. If you opt out of a section, you can always opt in later.
Each section is timed independently and you can take each section at your own discretion and pace.
Each question within a section is timed independently and varies from 1 to 7 minutes depending on the complexity of the question.
You can take breaks between sections, or do the whole test in one sitting: we leave that to you!
Have a pen/pencil and paper handy for your calculations and working.
Questions are typically worth 5 points. If a question has potentially multiple answers you can select, it will say “check all that apply”. You will receive partial credit for any correct option you have chosen. However, points will be deducted for wrong answers!
Once you complete all sections you will be asked to report your career aspirations, so we can recommend relevant AI roles.
Use Google Chrome to avoid technical issues.