abstractR

We were developing solution for one of the major auction sites in the EU region. One of the asks of the client was a service that would allow comparing auction items/auction prices from various other market places in the near realtime fashion. This was not part of the original specification, so we ended up looking for subcontractor that could deliver some kind of distributed crawler engine, so we could focus solely on the interpretation of the data in the original system. At first, we did approach local subcontractor we had past experience with, but unfortunately the cooperation failed in early stages as they did not have enough experience with the cloud platform our customer was using. After this experience, we put an add to local b2b marketplace, where we asked specifically for someone with previous experience on the platform and picked abstractR (mainly because of their review and the fact that they were active contributor on the forum associated with that marketplace). The requirements were one control node with simple web interface, where administrators could easily add new websites to crawl, the crawled content would later go to through various user defined processing pipelines. One of the requirements was to be able to dynamically spin new crawling nodes on the cloud platform used by the customer. Technologies used: python (implementation) celery (work distribution) redis (bloomfilter) lxml (html processing) elasticsearch (storage of processed data for later querying) They completed the project on time. Budget was exceeded by a little bit, but that was per our previous mutual agreement. Time was our most important constraint. Asynchronous architecture abstractR came with shown very good results and allowed to save on some hardware budget. It was enough to meet only once a week in the local coworking space. When we had any additional questions, we usually used IM to clarify the details in real time. We used shared git repository, so we knew exactly what was going on for the whole time. The flexibility and communication. Although the company was very small at the time (I communicated directly with Mr. Pylypenko), he was able to respond even outside of regular business hours. After the experience with the previous subcontractor, it was clear from the beginning that Mr. Pylypenko had actual previous experience with distributed computing, so this made the communication very easy.

"They went above and beyond to find the best solution and worked very closely with our product team." FEB 11, 2021 Quality: 5.0 Schedule: 5.0 Cost: 4.5 Willing to refer: 5.0 Feedback summary: abstractR developed a key fucntionaility that the CRM platform has been able to utilize for years. The team documented their work properly which allowed the client's internal teams to streamline support even after the engagement. They were communicative, organized, and overall easy to work with.