SoftwareSideHustles

For software engineers

Your coding skills could earn you more.

Discover flexible AI training opportunities where developers review code, evaluate responses, and help improve AI models.

  • Remote opportunities
  • Flexible project work
  • Potential additional income

Independent guide. Not affiliated with any AI training platform. Availability, eligibility and pay vary.

Try it · 2 minutes · no sign-up

Can you spot the stronger code?

An original practice task using the code-review skills AI evaluation draws on. Try it, read the explanation and decide what to practise next.

Write containsDuplicate(nums) that returns true if any integer appears more than once in nums, otherwise false.

Code A · javascript
function containsDuplicate(nums) {
for (let i = 0; i < nums.length; i++) {
for (let j = i + 1; j < nums.length; j++) {
if (nums[i] === nums[j]) return true;
}
}
return false;
}
Which implementation is better under these constraints?

The upside

See what the hours add up to

Drag the slider to set your hours. The figures are a simple illustration — real pay depends on the platform, your skills and current demand.

Illustrative earnings calculator

10h
1h40h

Illustrative hourly rate

At $50/hour for 10h a week, that's roughly

$500
per week
$2,167
per month
$26,000
per year
Illustrative earnings at 10 hours per week across different hourly rates
RateWeeklyMonthlyYearly
$25/h$250$1,083$13,000
$40/h$400$1,733$20,800
$50/h$500$2,167$26,000
$75/h$750$3,250$39,000

Hypothetical illustration only. It multiplies the numbers you enter; it is not a prediction. Rates, eligibility, project availability and actual earnings vary by platform, project, location and performance, and work is not guaranteed. Figures are simple arithmetic (rate × hours × weeks), hypothetical, and before any taxes or fees.

Want to earn this much?

Find the platforms that fit your skills — or build them up first.

The work

What technical AI training actually is

Frontier models are taught by developers who can tell good work from bad. Here's the work, in four parts — each with a free example.

  • Judge and compare code

    Decide whether an AI's code actually meets the brief, and which of two solutions is stronger — on correctness, complexity and the edge cases it forgot.

    See an example
  • Rank and label responses

    Place two answers on a fine-grained scale and calibrate “slightly” versus “strongly” better. Your consistency is what trains the model.

    Learn more
  • Check reasoning & instructions

    Spot explanations that sound right but contradict the code, and the requirements no test covers. Often the written justification is the deliverable.

    Learn more
  • Evaluate behaviour

    Rate honesty, calibration, safety and fairness from the evidence alone — trait by trait, without guessing beyond what's shown.

    Learn more

The journey

How it works

  1. 01

    Explore

    Understand the work

    See how developers review AI-generated code, compare solutions and explain their decisions. Start with the guides before choosing a platform.

    Know what the work involves

    Read the guides
  2. 02

    Find your fit

    Build your shortlist

    Use the questionnaire to compare your background and location with published requirements. Read each provider’s guide and check current eligibility.

    A shortlist to investigate

    Take the questionnaire
  3. 03

    Prepare

    Practise the core skills

    Try original code-comparison and code-review exercises. Learn to test edge cases and write specific, evidence-based feedback before applying.

    Practice with explanations

    Try free practice
  4. 04

    Apply

    Apply on the provider’s site

    Follow the provider’s application and complete any assessment independently. Review times vary; acceptance and access to projects are never guaranteed.

    An independent application

    Compare application routes
  5. 05

    Start working

    Learn the project’s guidelines

    If accepted and offered work, complete onboarding and read the task rubric. Use optional tutoring for general code-reading and reasoning skills, with original or public examples.

    A repeatable review method

    Get one-to-one tutoring
  6. 06

    Keep improving

    Build a feedback habit

    Reflect on the skills you need to strengthen, practise between projects and track your time. Optional tutoring can help refine your explanations without sharing confidential work.

    Clearer, more consistent evaluations

    Work on your skills with a tutor

This is a typical sequence, not a promised schedule. Each platform controls its assessments, acceptance, rates and project availability. Tutoring is optional and never includes live assessment answers or confidential client tasks.

Opportunities

Platforms that hire for technical AI evaluation

Summaries are based on each provider's public pages, with sources and the date we last checked.

  • Mercor logo

    Mercor

    358 open listings

    Published pay, technical roles

    A marketplace matching experts, including engineers, with paid AI evaluation and benchmark projects.

  • DataAnnotation logo

    DataAnnotation

    133 open listings

    Published pay, technical roles

    Remote contract work evaluating and improving AI output, including dedicated coding roles.

  • Alignerr logo

    Alignerr

    5,588 open listings

    Published pay, technical roles

    An expert network, powered by Labelbox, that pays specialists to help train AI.

  • micro1 logo

    micro1

    Published pay

    Rates shown on each listing

    An expert network that recruits specialists — including engineers — to train and evaluate frontier AI models.

Not sure which platform to explore first?

Answer a short questionnaire about your background and we'll suggest platforms, practice and next steps. No account needed.

Find Opportunities That Fit Your Skills

FAQ

Straight answers

Pay, acceptance, schedules and what we earn — without the hype.

What is data annotation?

Data annotation is the work of labelling, rating or correcting data so that it can be used to train or evaluate AI models. For language models this often means comparing responses, checking facts and writing explanations of what is right or wrong.

Learn more
What does coding-focused AI evaluation involve?

Typical tasks include judging whether AI-generated code is correct, tracing it against test cases, comparing two solutions, spotting where a response ignored an instruction and explaining your judgement in clear written English. The exact tasks depend on the platform and project.

Learn more
Is the work remote?

The platforms we cover describe their work as remote. Some projects have location-specific eligibility, so check the listing for your country.

Learn more
How much can people earn?

It varies widely. Platforms publish rate ranges for some roles, but actual earnings depend on the platform, the project, your location, your performance and how much work is available. We do not publish earnings promises, and our calculator is a hypothetical illustration only.

Learn more
Does SoftwareSideHustles.com guarantee acceptance?

No. We are an independent guide. Each platform decides who it accepts and what work it offers. Nothing on this site — including our questionnaire, practice labs or tutoring — guarantees acceptance or income.

Learn more
Does SoftwareSideHustles.com receive referral commissions?

Sometimes. Some links to platforms may be referral links, and we may receive compensation if you are accepted or hired through them. Every referral link is labelled, and we also show ordinary application links. Referral arrangements never change how a platform assesses you.

Learn more
See all questions

Put Your Coding Skills to Work

Start with the free practice lab, or answer the questionnaire to see which platforms fit your background.

Cookies on this site

We use essential cookies to run the questionnaire, practice labs and sign-in. With your permission we'd also like to use analytics to understand which pages help people, and advertising cookies to measure our Google Ads. Rejecting never limits any feature. Cookie policy