General Talks ~ Tensorflow + Kubernetes + ML Pipeline




General Talks ~ Tensorflow + Kubernetes + ML pipeline

29 April 2019


Added 01-Jan-1970

Tentative agenda:

18:20 - 18:30 - Check the names at the entrance.

18:40 - Talks.

20:00 - Networking, drinks, pizza, sweets and snacks.


Speakers of the day (not necessarily in this order):

1) TensorFlow 2.0
Daniel 'Wolff' Dobson
Google Developer Programs Engineer

2) Title
Samantha Guerriero: Building an end-to-end ML pipeline for Object Detection

Computer Vision is one of the hottest topics in AI. Building state-of-the-art Computer Vision models is within anyone’s reach thanks to the high availability of open-source research, online support, user-friendly high level APIs and ready-to-use pre-trained models. But how do you put a Computer Vision model into production and serve it in a real life application?
Sam will walk us through an effective Object Detection built on GCP for one of Datatonic’s clients. The solution uses TensorFlow Object Detection API, ML Engine, Cloud Composer and Cloud Functions to identify stories on screenshots of news websites, automatically re-train the models and create annotation boxes for new data as it lands on GCS. This talk will give you a great example of how to effectively design an ML pipeline on Google Cloud.