Textkernel highlights: looking back at (the second half of) 2019
11-12-2019 • Categories: Annual review, Clients, News, TextkernelIn the spirit of good cheer and holiday wishes, we bring you an overview of all the great things Textkernel achieved in the last months since Summer break! In addition to the below solution-specific enhancements, we are always releasing various enhancements to our infrastructure to improve our platform speed and stability. We strive to bring our customers faster response times and improved uptime. We are very proud of what we have accomplished, both in the last six months as well as the first half year.
We’ve been hard at work to improve our Match service to bring you more relevant results, as well as the ability to match Italian and Spanish language vacancies and applicants. Here are the most noteworthy developments:
Improved matching excellence:
Whether you’re matching candidates to jobs or vice versa, our customers expect only the best results. To keep your results list at a top-notch level, we’ve made adjustments and improvements based on the latest research and customer feedback.Our improved query generation algorithm enhances approximately 20%-30% of the formulated Match queries, and significantly improves recall (finding as many relevant results in your index as possible) on about 10%-20%. These improvements are language-dependent, but provide an indication of our ongoing efforts to ensure query results are relevant and accurate.
Our ongoing updates are released on a continual basis, so you are likely to see some differences in the auto-generated match queries over time.
Match candidates in more languages:
In October we launched our Match! solution in Italian and Spanish! Matching candidates to jobs and vice versa has never been easier. Now you can generate a candidate search query based on the requirements listed within an existing open job requisition. Or, you can identify a list of suitable job vacancies based on the CV of an available candidate.
One of the key benefits of using Textkernel’s Match! service is that you can match jobs and candidates regardless of Match! language combination. For example, if you had a great Italian software developer and an open (hard-to-fill!) vacancy in the Netherlands, you could match both Italian CV document with job requisition requirements based on our technology.
And of course we offer more than just AI-powered matching. Here are some of the quality improvements we’ve made for our other solutions:
✓ Enhanced CV parsing accuracy: We’ve launched a number of parsing-related enhancements so you can process documents faster and more reliably than before:
- We’ve upgraded our parsers for Czech, Polish, Romanian and Slovak to Extract 4.0 (deep learning) technology. The outcome: Our parsing accuracy is 30-40% better for those languages! Learn more
- We’ve also made over 30 micro-improvements to our parser in other languages, improving our accuracy for multiple language offerings.
- We have upgraded our OCR software to ensure that scanned CVs are parsed better.
Looking to enrich your current database with additional classification fields?
✓ Data enrichments in CV & Job Parsing: ISCO and O*NET classifications for job titles are now available out-of-the-box for English, French, German, Dutch, Spanish, Portuguese, Japanese and Chinese.
✓ Geo-coordinates of cities can be added to the output of the parser. Contact your account manager for information on upgrading your account.
In a fast-moving jobs market, data quality is paramount:
✓ Jobfeed data quality improvement: we are enhancing the identification of staffing and recruitment agencies to enhance our reporting on posting organization types. This will help ensure that queries for direct employers returns only those really from this organization type and not re-posted vacancies on the employers’ behalf.
✓ Increased number of Jobfeed data sources: We’ve added approx. 1,700 new data sources to Jobfeed in the past 6 months. These are predominantly from the US and UK labor markets.
Learn more about Textkernel’s labor market intelligence.
In addition to the above solution-specific enhancements, we are always releasing various enhancements to our infrastructure to improve our platform speed and stability. We strive to bring our customers faster response times and improved uptime.
Wishing You a Festive Season!
The Textkernel Team