Looking for Annotators

Critical part of our system is the object detection and classification function. We are constantly making our machine learning models better through increasing amount of datasets we use for its teaching. The challenge with datasets is that we need to annotate them before they can be used, which means that we identify what kind of objects there are in the picture. This process is called annotation process and to carry out this work we employ freelancers. Freelancers in this case can be retired sea personnel, students or people working at sea. At the moment we can unfortunately only employ people in Finland through so called ligh entrepreneurship model which means that the freelancer will invoice us from the work carried out.

What kind of work is available

Work is flexible, work as much as you want, wherever you want, whenever you want.

The work is done on your own computer using Groke's annotation tool, which works with a Chrome web browser. A big screen and a good mouse make the job easier.

Groke can always be contacted via the slack channel, where you can also find other annotators and quickly get an answer to questions.

The salary is agreed with the annotator.

The work requires accuracy, good quality and diligence of annotation.

The employment relationship is carried out through light entrepreneurship mode like Eezy.

How annotation is taught

Small free-form interview and presentation of the annotation tool

• We'll tell you about the job in general.
• The annotator is taught how to use the tool, how to log in, how to draw bounding boxes and how to give the values as well as how AIS objects should be attached to bounding boxes.
• We also instruct you how to rhythm your work and track progress.

The annotator is provided with documentation of the desired annotations:

• What to annotate
• How to annotate
• Annotation categories
• Other object attributes

The annotator is given a small test task to perform. The test task is checked by GROKE and after that the annotator can start the work.

If intrested contact our machine learning engineer Ville Vianto:

[email protected]