Upwork is hiring a manual data labelling of EU policy reports - Contract to Hire

Job title:Manual data labelling of EU policy reports

Skills: English, Excel

Billing type: to be agreed

Timings: Start - ASAP duration - 1 month, hopefully permanent

Location: Work from anywhere

Hiring process: 30 mins interview

About the role

We are looking for someone to manually label unstructured data.

We are training an AI model that can determine the stance (‘in favour’, ‘against’) about issues mentioned in EU policy reports.

The AI model analyses a set of topics in a report and attaches a label ‘in favour’ or ‘against’. We want you to check the outputs of the AI model and label if the output of the AI model is correct or not. You will do this by reading the sentence or the paragraph in which the topic appears and interpreting that sentence or paragraph. For example, if the AI model says that on page 6, the report is ‘against’ (label) ‘chemical recycling’ (topic), then you need to check page six and see if this is true, by interpreting the text yourself. Here’s an example of an EU policy report which the model will analyse: https://op.europa.eu/en/publication-detail/-/publication/05a3dace-8378-11ea-bf12-01aa75ed71a1

You will label the data in MS Excel or Google Sheets. The reports will be in PDF.

We expect you to label 1,000 - 2,000 entries. We expect this will take around 5-10 days to complete.

If you do this well, and you are keen to progress, you can then take on more interesting projects in many areas such as new product development, BE / FE development, working with our data scientists to do data analysis / visualisation, develop NLP models, etc.

Qualities we are looking for

You need to be pretty good in English and have a good feeling of semantics in general. You must have an eye for detail.

About us

Hi there, we are www.lobium.ai. We make data work harder for Public Affairs.

We are building a B2B portal that helps professionals in the fields of Politics, Policy and Public Affairs receive valuable and actionable intelligence so that they can engage more successfully in legislative processes. For example, using NLP, we have built a stakeholder-issue analysis tool that discovers and plots stakeholders’ positions related to a legislative initiative. And we have lots of other really cool ideas on our roadmap.

We use a modern, exciting tech stack and we are proud of the relaxed, innovative and collaborative environment we have created. We write our backend in Python and our Front End in React. We use Neo4J as our graph database management system. We are solving unique and fascinating challenges, and we have a great team of super-bright people. Lobium team members are based all across the globe and we hope to expand further.