Google Developer Groups DevFest — 2017

Saturday, October 14th, 2017 | HootSuite, Vancouver

The Google Developer Groups (GDG) DevFest in Vancouver is an all-day developer conference that had 2 tracks: an App Track and a Cloud Track.

  • In the App Track, there were sessions covering technologies for application development such as Flutter, Firebase, Android and Progressive Web Apps, as well as how to scale applications with backend services and cloud.
  • In the Cloud Track, there were talks regarding web-backends, Kubernetes, internet-of-things, as well as data processing, dev-ops best practices, and machine learning using Tensorflow.

I attended the Cloud Track.

Morning stretches before separating into the 2 tracks
Coffee is a must have
Swag is always good

Deploying Applications to Cloud

Drew Ogryzek, Lead organizer of HackerNest Vancouver

GO, Docker, Codefresh, Kubernetes, and Google Compute Engine (GCE) were used to demonstrate how to deploy applications to the cloud in 30 minutes. Drew’s tutorial is available at

Infrastructure management on GCP

Dana Hoffman, Software Engineer for Google Cloud at Google

The premise of Terraform, an open source infrastructure management tool, is to create infrastructure as code which can be edited, reviewed, and versioned. Combining with Google Compute Platform (GCP), Dana demonstrated how to do load balancing in a variety of scenarios. A version of her demo is available at

Pizza for lunch. Is 101 boxes too much?
Time for the next session

3 Lightning Talks

Be unreasonable. Be responsible.
— Hussein Hallak, GM of Launch Academy
We welcome creators and enthusiast.
— Melissa Dex Guzman, Vancouver VR community lead
Always be iterating.
— Alim Jaffer, VP Product at Heroic Labs

Social Media Data Analysis

Kiana Mohseni, Senior Director of Development at HootSuite

HootSuite, a platform for managing all your social media in one place, is targeted for marketers and those interested in measuring, enriching, and increasing your social media presence.

Industry is all about maximizing the ROI (return on investment)
No machine learning talk is complete without a shoutout to minimizing the loss function by gradient descent

TensorFlow and Machine Learning

Vikram Tiwari, Google Developer Expert for Google Cloud
In lieu of Josh Gordon, Developer Advocate for Tensorflow at Google

After giving a general overview of machine learning, mentioning there is no active development on Theano after the 1.0 release, and that TensorFlow is the way to go, Vikram demonstrated the use of pre-existing TensorFlow models retrained on the images of 3 types of noodles to predict the noodle type of a new image. His code is available at

Lots of leftover pizza

Listening to talks, getting inspired, and finding out what other people do is not a bad way to spend a Saturday. Overall, it was a pretty satisfying and interesting day.