GoSF - Building Data Pipelines (times 2) + Machine Learning and Go

Wednesday, May 16, 2018 - 18:00
San Francisco

6:00 Networking | Food | Drink
6:30 Speakers
• Talk 1: Building Resilient Data Pipelines in Go (Grant Griffiths, GE)
• Talk 2: Machine Learning and Go
• Talk 3: Go to the Cloud with Kafka! Motiv Ring's Migration from Client to Server-Side Logic


About Okta:
Okta is an API service for storing user accounts and handling authentication and authorization. Okta makes it easy to add user registration and login to your apps, as well as things like multi-factor authentication, social login, single sign-on, and more.

About Shift:
Shift is a better way to buy and sell cars. We are leveling the playing field for car buying by pairing a technology-first approach with superior customer service. Our platform is built with Go, React, and Swift, and uses machine-learning to accurately and transparently price cars. Our end goal is to build a more trustworthy marketplace and make it really simple for anyone to buy, own, and sell their car.


Talk 1: Building Resilient Data Pipelines in Go

Grant Griffiths from GE will provide an introduction to writing resilient data pipelines in Go using Kafka and Cassandra. He will cover how to build a very basic data pipeline, run it in a Kubernetes cluster, and write tests to simulate downtime of various components.

At the end of the talk, Grant will go over an example of how GE Digital uses these concepts to build a data pipeline for parsing and ingesting larges volumes of time series data.

About the Speaker
Grant Griffiths is a Software Engineer for GE Digital on the Predix Cloud Engineering team. He works on building data services in Go for Predix, a cloud platform for storing and analyzing Industrial IoT data. In addition, Grant also runs the Go User Group for GE, hosting internal talks and demos across the company.


Talk 2: Machine Learning and Go

Did you know there are more variations of cars built in the last 10 years than there are used cars for sale? Not to mention every car has a unique backstory. In this talk, Andrew Carman will discuss the challenges of used car pricing with machine learning and how we deployed them into our production Go environment.

About the Speaker
Andrew Carman is a software engineer and amateur triathlete from San Francisco. He studied CS and math at Harvey Mudd College and has spent his career at working at early-stage startups building infrastructure, data products, and APIs. He tries to live every day like it’s Pi day and has been told his celebrity lookalike is Hiccup, from How to Train Your Dragon.


Talk 3: Go to the Cloud with Kafka! Motiv Ring's Migration from Client to Server-Side Logic

Dan Post from Motive will provide another view on data pipelines in Go based on Kafka. In the course of his talk he will describe their journey in designing and building a data pipeline to replace a bunch of client-side logic. He will explain their unique (or not so unique) requirements and performance needs and how they translate into the eventual design and implementation.

About the Speaker
Dan Post is the Director of Engineering at Motiv and is responsible for the development of the Motiv Ring, a fitness tracker that fits in a waterproof titanium ring. Prior to Motiv, he held senior engineering roles for 12+ years at Square, Apple, and Intel.


301 Brannan Street, 1st Floor