RSS feeds in the directory: 2794
Added today: 0
Added yesterday: 0
Machine learning has become a fashionable term, but when working with large amounts of data it has for many years been a vital necessity. Yandex processes over 200 million queries daily! Once on the Internet there were so few sites that the best of them were placed in the directory and now for the relevance of page links in the results to answer complex formulas, trained on new data. The task falls on the so-called conveyors, regular processes, training and controlling these formulas.
Today I want to share with the community Habra our experience of creating a computing platform Nirvana, which, among other things, used for machine learning tasks.
Nirvana — specialized cloud-based platform for managing computing processes, where applications are run in the order specified by the user. In Nirvana has the process description, links, blocks, processes and related data. Processes are documented in the form of asynchronous graphs.
Nirvana for the solution of computational problems used by developers, analysts and managers of different departments of Yandex — because not everything can count on my laptop (and why — we will tell you at the end of the article, when we move to examples of the use of Nirvana).
, We describe the challenges faced when using previous solutions, will describe the key components of Nirvana and the reasons why the platform was chosen such a name. And then look at the screenshot and move on to the tasks for which useful platform.
Read more →
[Sandbox] Development of highload WebSocket service 12.03.2018 at 15:51
MathOps mathematics or in monitoring 12.03.2018 at 12:00
In VTB changed the paradigm of data access 12.03.2018 at 09:47
[Translation] what about the "nihilists in infobezopasnosti" 12.03.2018 at 08:29
Is it easy to address the conference for the first time? 12.03.2018 at 07:33
FINTECH-digest: banks without offices are becoming more popular 12.03.2018 at 07:03
To remember them all. Expectation and reality in IT: 2017 vs 2018 12.03.2018 at 03:17
[Translation] the Lessons that can be drawn from the code base of LLVM/Clang 12.03.2018 at 01:38