The Client

Our client is a European company that delivers natural language and search based analysis on textual data. Their technology enables users to just point to their data and start asking questions or search to get interactive results from data.

The Challenge

Our client wanted to train their model to extract named entities in business documents. For the algorithms to understand the information they are being fed, they must be fed large volumes of relevant annotated training data. The challenge was linking relevant text strings and metadata labeling. Our client wanted to save the time of their data scientists and get high quality annotated data at scale.


The Solution

Since we were dealing with huge passages of text with technical language we formed a team of experienced annotators lead by an industry expert with an end to end customized workflow:

  • Alignment of text annotation tools and processes
  • Structured development milestones
  • Two-step production and QA annotation workflows

The Results

A team of 15 annotators worked on the task for a month and we were able to annotate some tens of thousands of lines of textual data. Our efforts freed up the data scientists’ time to work on their core technology instead of spending time on data annotation and QC. Our work also improved the performance of the models by a significant amount.

Interested in a free pilot?

Contact us with your requirements and we will set up a team to wok on your free pilot project. No commitments on your side.

Contact us