-
Notifications
You must be signed in to change notification settings - Fork 0
About
My name is Martin and I am the author of NGRAMS.
I am a backend developer and like to build software that deals with large amounts of data. I do things in C++ because it's fast — at least at runtime, not necessarily at development time. I also like web programming to some extent thanks to TypeScript. When I should name my most important software development principle it would be Keep It Simple. I received a master's degree in computer science in 2012 from the Bauhaus-Universität Weimar (Germany).
Email me at [email protected]
NGRAMS is my third implementation of a search engine of this kind.
2019 - today — ngrams.dev
Dataset: Google Books Ngram Dataset v3
Size: 23 TB compressed, ~230 TB uncompressed
Backend: C++20 for core app, uWebSockets for REST API server
This thing has been released mid April 2023.
2015 - today — phrasefinder.io
Dataset: Google Books Ngram Dataset v2
Size: 7 TB compressed, ~70 TB uncompressed
Backend: C++14 for core app, Boost Beast for REST API server
This thing will be discontinued by the end of 2023.
2007 - 2013 — netspeak.org
Dataset: Web 1T 5-gram Version 1
Size: 25 GB compressed, ~75 GB uncompressed
Backend: C++03 (later C++11) for core app, JNI, Java Servlet for REST API server
This thing started off as my bachelor thesis.