We held CloudSearch Meetup at Amazon Meguro office in Tokyo.
In March 2014, Amazon CloudSearch launched a lot of new features. And now, CloudSearch team is in Tokyo on business. So it is the best timing to hold a meetup here.
– CloudSearch Overview, Technical deep dive by Jon Handler(@jon_handler)
CloudSearch Solution Architect Jon Handler gave a presentation about Amazon CloudSearch and features.
[slideshare id=34754376&w=427&h=356&fb=0&mw=0&mh=0&style=border:1px solid #CCC; border-width:1px 1px 0; margin-bottom:5px; max-width: 100%;&sc=no]
In Q&A part, there were a lot of pragmatic questions like Having customer original dictionary / Language detection / N-Gram tokenizer / Realtime indexing and so on.
That kind of feed backs are really precious to CloudSearch team. Thanks a lot!
– Expectations for CloudSearch by Osuka-san(@mosuka)
Osuka-san is the contributor of Apache Solr and he gave CloudSearch some expectations. He has been focusing on Search systems for a long time so his saying was very helpful to understand how hard to manage Search infrastructure by ourselves and using CloudSearch is beneficial.
[slideshare id=34742755&w=427&h=356&fb=0&mw=0&mh=0&style=border:1px solid #CCC; border-width:1px 1px 0; margin-bottom:5px; max-width: 100%;&sc=no]
He told us some detailed Japanese matters.
“こちらは社長室長谷川さんです” – There are two meanings in Japanese.
1. This is Hasegawa-san, Member of President’s Office.
2. This is Tanigawa-san, Executive secretary of President’s Office.
Even for Japanese people, we need to understand the “context” to tokenize. Yes, Japanese Search is not easy.
Aside from that He proposed using Machine Learning technologies to solve Search quality issues and AWS has variety of services to handle that. So having connectors to that services are nice.
And nowadays, make the information searchable quickly is the demand so having Kinesis connector to manage real-time is really welcome.
– Impression of Using Cloudsearch(@yoshi0309)
Yoshida-san is very skilled Search engineer and he wrote some blog post about CloudSearch. I was lucky to have Yoshida-san in our meetup 🙂
He gave us some topic that he did try to use. Not only explanation but this kind of DEMO-ish presentation is awesome!
Maybe the cost would be much cheaper than his estimation but CloudSearch can appeal cost effectiveness compare to other Search solutions.
And we are looking forward to having CloudSearch functions in Manifold CF that Yoshida-san has done pull request!!
– SnapDish usecase by Kiyota-san(@kiyotaman)
We appreciate that Kiyota-san always shares SnapDish CloudSearch use case kindly. Thanks to Kiyota-san, CloudSearch can appeal that Japanese AWS user has already used CloudSearch in popular service.
[slideshare id=34743444&w=427&h=356&fb=0&mw=0&mh=0&style=border:1px solid #CCC; border-width:1px 1px 0; margin-bottom:5px; max-width: 100%;&sc=no]
For fine-tune, he has done Japanese Tokenize himself(means outside of CloudSearch). CloudSearch team understands that and will do some work for it.
Everyone was hungry because the meetup was held in late at night and his SnapDish slide made us more hungry 🙂
He leads Python Conference in Japan(called PyCon). If you are a Pythonista why not do join? https://pycon.jp/2014/
We did Drinkup after meetup. CloudSearch team and Japanese Search engineers had nice discussions 🙂
Osuka-san and Vivek
My Search SENSEIs(we worked at the same campany)!
Thanks for coming, guys!!
Thanks for helping me Yifeng and Yasukawa-san to organize the event!!