> ## Documentation Index
> Fetch the complete documentation index at: https://kb.micro1.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Types of Exercises

> Different challenges so you can assess your candidates knowledge.

### Coding Exercise

A [Coding Exercise](./coding-exercise) is a timed technical challenge designed to assess a candidate's programming skills in a real-world scenario.
Candidates have **up to 25 minutes** to complete the task using a **simple, browser-based IDE** that allows them to write, edit, and **run their code directly** — no setup required.
This format is ideal for evaluating problem-solving ability, code quality, and familiarity with core language features.

Supported languages include:

* C
* C++ (Clang or GCC)
* C#
* Go
* Java
* JavaScript
* PHP
* Python (v2 or v3)
* Ruby
* Rust
* Swift
* TypeScript

<Tip>
  If you need a different language not listed here, check with our team.
</Tip>

### Custom Exercise

A [Custom Exercise](./custom-exercise) allows you to create tailored evaluation tasks by defining **your own instructions**.
Candidates will respond either by typing into a provided **input textarea** or by **recording audio**, depending on the format you choose.
You can also set a **time limit** for completion to simulate real-world constraints and maintain consistency across interviews.

Examples of what you can assess with Custom Exercises:

* **Script reading** *(audio)*: Evaluate pronunciation, tone, and fluency
* **Writing an email** *(typed)*: Assess clarity, professionalism, and grammar
* **Answering a chat message** *(typed)*: Test responsiveness, empathy, and communication style

### Data Annotation Exercise

A [Data Annotation Exercise](./data-annotation-exercise) is designed to evaluate a candidate's ability to assess and refine AI-generated outputs.
You start by defining a **subject** and a **context**, and our AI will automatically generate a relevant annotation task.
The candidate's role is to **evaluate the AI's responses** based on your instructions—identifying errors, suggesting improvements, or validating correctness.

This format is especially valuable for vetting domain experts who will contribute to **training and improving AI models** through high-quality feedback.

* Define a **subject** and **context**
* AI generates a tailored annotation task
* Candidate evaluates AI-generated answers
* Ideal for building expert teams for **AI training and validation**