Accelerating data insight for a better work life

NEW WORK SE thrives on fast, flexible access to rich data, customizing service and recommendations for employers and professionals

+ show more
Case study

In the not so distant past, jobs used to be scarce and employers were free to choose new hires from an abundance of contenders. How times have changed: The digital transformation has revolutionized the labor market. Today, people with the right talent tend to be in short supply. Extending the reach of working people and organizations, NEW WORK opens a new world of opportunities for seekers and providers of skilled employment across all industries. NEW WORK makes it its business to focus on the success of people and companies. The brands owned by NEW WORK include XING, kununu, HalloFreelancer, InterNations, Prescreen, and Honeypot.

  • Challenge

    Intelligence for faster, more inclusive fits

    NEW WORK continuously trains its analytic models to help job seekers optimize their presentation such that they can quickly be found by recruiters. For employers, NEW WORK’s matching software automatically analyzes the experience, skills, and job history of contenders, queries their availability, and notifies companies with a shortlist of the most suitable candidates.

    NEW WORK thus empowers employers to fill vacancies much faster than conventional recruiting processes would.


    The company is also committed to inclusion and diversity. Not everyone has equal access to jobs, but NEW WORK believes that artificial intelligence can help organizations take an unbiased approach to finding people with the right skills while at the same time creating mixed teams, maintaining gender parity, and fostering inclusion.


    Instant data access for fast analytics

    NEW WORK provides its subscribers—individuals and organizations—with actionable recommendations driven by algorithms. To train these algorithms, the company’s analysts, data scientists, and developers need relevant data for analysis. The challenge for NEW WORK was to enable these specialists to extract the right data quickly and efficiently from a vast lake of raw data. This data lake includes structured data such as member and company profiles. It also includes semi-structured and unstructured data that comes in various formats ranging from JSON, to ZIP files, to photographs.


    Traditional solutions such as Network File System (NFS) stores and relational databases were not addressing the company’s requirements in terms of flexibility, performance, and cost-effectiveness. The business intelligence (BI) team at NEW WORK concluded that none of the open source data management software in the marketplace were enterprise-ready.

  • Solution

    Flexible, fast, and scalable data platform

    The only software solution that fully delivers on NEW WORK’s expectations is HPE Ezmeral Data Fabric, formerly MapR Data Platform. “We need a stable, massively scalable and easy to manage environment for keeping huge amounts of raw data accessible for analytics,” says Eren Akca, Senior Business Intelligence Systems and DevOps Engineer at NEW WORK. Also, HPE Ezmeral Data Fabric is the only software solution that provides the multitenancy capability NEW WORK needs to keep customers’ data segregated in line with data privacy legislation.

    The company’s data lake holds a very large number of small files that used to slow down performance. HPE Ezmeral Data Fabric overcomes performance issues by providing a logical layer that integrates all of NEW WORK’s data sources. The BI team at NEW WORK finds the file system provided by HPE Ezmeral Data Fabric straightforward to use. The file system handles a wide variety of workloads, including snapshots and even backup mechanisms.

  • Benefits

    Smooth start, low overhead

    NEW WORK’s BI team started with a fairly small deployment of HPE Ezmeral Data Fabric, involving only a few data scientists. This approach helped the company keep hardware, administration, and maintenance costs low.

    Today, HPE Ezmeral Data Fabric serves NEW WORK as a powerful, scalable, and cost-effective data management software solution, enabling the company to ingest, store, manage, and analyze rapidly growing volumes of data—all in one place. “Admin and maintenance would usually take up 20 to 30% of our team’s time. HPE Ezmeral Data Fabric reduces this to less than 10%. For example, HPE supplies maintenance services for this software in convenient bundles. That frees up time, so we can focus on services to analysts,” Akca explains.


    Straightforward and efficient

    Providing raw data to the company’s specialists for quick exploration and analyses is a straightforward exercise for the company’s BI department. “We load raw data to HPE Ezmeral Data Fabric, perform quality assurance checks, fix issues, implement improvements, and then generate a data pipeline for analytics,” Akca says. Self-service access to the data lake for instant exploration of use cases and analyses is well received by NEW WORK’s analysts, data scientists, and developers. A key benefit provided by HPE Ezmeral Data Fabric is the ability to run multiple use cases on a single cluster instead of having to rely on a separate cluster for each use case.


    New employees at NEW WORK find HPE Ezmeral Data Fabric intuitive to use. In the past, it would have taken roughly one day to activate an additional data source. In a recent example, it took only ten minutes to stream data from a new source into the HPE Ezmeral Data Fabric file system and provide access to the team.


    Cutting the cost of data provisioning

    At NEW WORK, HPE Ezmeral Data Fabric takes care of the heavy lifting of raw data. If required for a given use case, the BI team will cleanse and aggregate the associated data and then move it to the data warehouse. Such data typically remains in the data warehouse for several months, after which it is returned to the data lake.


    Keeping raw data in the data lake saves the company a considerable amount of money. “The cost of storing huge amounts of raw data in our data warehouse would be forbidding,” Akca points out. “Also, keeping the vast bulk of raw data in the lake helps us prevent any potential overrun and performance impairment of our data warehouse.”


    The BI team plans to deploy the HPE Ezmeral Container Platform in the near future. Designed for both cloud-native and non-cloud-native applications, this platform will support NEW WORK to manage data fabric clusters, new software, and open source tools to exploit the raw data. The BI team sees HPE Ezmeral Container Platform as a powerful and logical complement to HPE Ezmeral Data Fabric.

Download the PDF