Get Started
Hunting for you, 24x7Get the iPhone app

Your hunt runs in your pocket, and tells you the moment a strong match opens.

Download on theApp Store
TolokaPosted in November 2026

Freelance ML Researcher (Part-Time)

European Union (Remote)RemotePart-time55 – 60 an hour
Who we'd reach at TolokaFound and reached by Power Apply
Best leadHiring manager

Leads the team this role joins

Recruiter

In-house, owns the opening

Posted this job

Put the opening up

Power Apply

From your own Gmail and LinkedIn.

About the role

This is a part-time freelance machine learning research role at Toloka, fully remote from within the European Union. It suits an engineer who has hands-on experience fine-tuning large language models and wants to turn experiments into published research. You will design and run short, self-contained projects that reproduce and test post-training experiments on LLMs using Toloka's platform, then publish each one as a blog post under your name. Toloka supplies the platform, the GPU compute, and a researcher from its ML team who reviews your plan and results. The company works with frontier AI labs and enterprises on training data built from expert and crowd contributors, and its ML team post-trains open-weight models so smaller models can match larger ones on specific tasks. The team wants to use this role to publish more of its post-training work and to make Toloka known for that research

What you’ll do

  • Reproduce recent papers or posts that gain attention, add your own ablation, and report which results hold up.
  • Point a coding agent at a training problem, let it run many experiments, and document what it tried, what worked, and where it stalled.
  • Run small head-to-head comparisons on one well-defined task across Toloka's platform and alternatives, measuring quality, cost, and effort.
  • Write each project up as a blog post that covers the full path, including failed experiments and how you reached the final result, with help from the team on editing and charts.
  • Report platform problems and feedback directly to the engineers who build it, as one of the first outside users.

What they’re looking for

  • Hands-on experience training or fine-tuning LLMs with PyTorch, the Hugging Face stack, LoRA, or supervised fine-tuning.
  • The ability to take a vague research idea to a result independently, design the experiment, and separate real effects from noise.
  • Current working use of coding agents such as Claude Code or Codex.
  • Clear technical writing in English.
  • A link to prior work such as a blog post, paper, or repository.

Nice to have

  • Experience with reinforcement learning for model training.

What’s on offer

  • Freelance B2B contract paid through Deel at $55 to $60 per hour.
  • Part-time work of about 40 hours per month with a flexible schedule, and the option to concentrate hours into a one or two week sprint when a topic is active.
  • Fully remote work from the European Union.
  • An initial three-month engagement with the option to extend.
  • Access to Toloka's post-training platform and GPU compute for your experiments.
  • Publication of your research on the Toloka blog under your name.
  • Direct collaboration with the researchers and engineers who build the platform.

Skills

Questions about this role

Is this role remote?

Yes. The role is fully remote and open to candidates in the European Union.

How much does it pay?

The pay is $55 to $60 per hour under a freelance B2B contract paid through Deel.

How many hours per week or month is the role?

The role is part-time at about 40 hours per month, with a flexible schedule.

How long is the contract?

The initial engagement is three months, with the option to extend.

What skills are needed for this role?

You need hands-on experience fine-tuning LLMs with PyTorch, the Hugging Face stack, LoRA, or SFT, plus clear English technical writing and regular use of coding agents.