ML Engineer (MSc. or PhD)

What you'll do

The tasks are ML engineering rather than ML theory. A representative one asks you to implement or modify a model, wire up a training or inference workflow that reproduces cleanly, squeeze memory use or throughput, chase down a numerical or system-level failure, and show that the result meets objective correctness and performance targets. You build these problems, solve them and review others, all through Python and the command line, and micro1 describes the role as very coding-heavy with coding agents expected in your workflow.

Remote contract work at roughly 15 hours a week, on hours you pick.

Who qualifies

• A master's degree or PhD in Computer Science, Machine Learning, AI, Applied Mathematics, Statistics, Engineering or a similar quantitative field (unlike some siblings in this family, the degree is listed as required, with no alternative route)
• Strong professional or research experience in machine learning
• Practical Python and hands-on use of coding agents
• Strongly preferred: time at an established tech company, AI lab, research organisation or other recognised engineering environment; exceptional open-source or academic work may substitute

The detail panel titles the role "Senior ML Engineer", which matches the bar above.

What it pays

micro1's published band is $60–120/hour, paid per task that meets the project specification, with minimum submission requirements. That is a lower band than micro1's existing ML Engineer listing at $100–150/hour, a separate posting, so if you qualify for both it is worth comparing the scope before choosing.

Worth knowing

Good:

• Systems-flavoured ML work (performance, reproducibility, debugging) that suits people who ship models rather than only publish them
• 100 openings, flagged new and high demand
• Global and fully remote, no country restriction stated
• Short screening: questions, a roughly 30-minute AI interview, a hiring manager review

Less good:

• The degree is a hard requirement here, so strong self-taught engineers should look at sibling listings instead
• Paid per accepted task; a debugging rabbit hole is unpaid time
• Minimum submission requirements are mentioned but not defined
• micro1 expects a first task within 24–48 hours of onboarding

About this listing

We read this ad on micro1's signed-in jobs board on 18 October 2026, three days after it was posted. It belongs to a set of "(MSc. or PhD)" expert listings micro1 published together on similar terms. The customer, the project length and the payment schedule are not given. See micro1.