AWS re:Invent 2018

Latest Episodes:

STP07: Kabbage's Long Strange Trip to Data Democratization Dec 01, 2018

Show notes

In 2016, small business financing company, Kabbage, began forging a path to migrating its aging data warehouse infrastructure to something new. In this talk, Clint Hill, head of architecture and platform solutions, and David McGowan, VP of engineering at Kabbage, share how the Atlanta-based company moved away from a static relational table, muddled through a project with incomplete vision, little to no design, and a bespoke data management system, and ultimately found its way to a solution comprising several AWS Managed Services-ultimately creating a method that provides user flexibility and solves real business issues.

STP08: How Robinhood Used AWS to Make a Self-Service Data Platform Dec 01, 2018

Show notes

Data and data-based decisions are at the core of everything online brokerage Robinhood does. In this talk, we dive deep into how Robinhood moved from a world where multiple systems grew to create an unwieldy data architecture, and we discuss how the company used AWS tools, such as Amazon S3, Amazon Athena, Amazon EMR, AWS Glue, and Amazon Redshift, to build a robust data lake that can operate on petabyte scale. We also discuss design paradigms and tradeoffs made to achieve a cost-effective and performant cluster that unifies all data access, analytics, and machine learning use cases.

STP09: How iflix Uses AWS Elemental Services to Redefine the Future of TV Dec 01, 2018

Show notes

At iflix, we want to lead the internet entertainment revolution in emerging markets by redefining television for billions of people. While not everyone who wants great content may have a credit card or be able to afford pricey subscriptions, most do have smartphones, and that's where iflix comes in. In this talk, we address how we've had to be extremely scrappy to successfully serve streaming content to our millions of customers, and we discuss how we leverage various AWS services to accelerate our growth-the most recent being AWS Elemental MediaLive & AWS Elemental MediaPackage, which power our live TV option.

STP12: Crypto 101/201 Dec 01, 2018

Show notes

Blockchain technology has the potential to dramatically change the financial landscape. Kristen Stone, the product manager on the coinbase cryptocurrency payments team, and Jake Craige, the team's technical lead, join us to discuss the evolution of the blockchain industry and its impact on our day-to-day lives. They dive deep into some of the inner workings of the technology, and they discuss the different types of blockchain models, smart contracts, consensus protocols, and what's on the horizon.

STP14: How Coinbase Builds Out Its Blockchain Infrastructure Dec 01, 2018

Show notes

Coinbase has some fairly unusual security and infrastructure requirements. One of these requirements is that every server in our infrastructure is both ephemeral (30 days) and immutable. The deployment process for most applications is fairly straightforward: 12-factor apps are blue/green deployed behind a load balancer. Blockchain nodes, however, present a difficult problem: How do you blue/green deploy a server with 1 TB on disk? In this session, we discuss how we solved this problem in a blockchain-agnostic fashion using a new project, called Snapchain.

STP15: How the Cloud Helps Nubank Support Millions of Daily Customers Dec 01, 2018

Show notes

Nubank, a Brazilian financial technology company, is empowering customers with a 100% digital banking experience. Born in the cloud, Nubank does this by enabling its engineering teams to create scalable software that can quickly adapt to the changing needs of a complex market and by fostering a culture of microservices, container-based applications, observability, and immutability. In this talk, Nubank engineers Alexandre Cisneiros and Diogo Beato discuss why they treat infrastructure as a software engineering problem, and they address how AWS services, like AWS CloudFormation, Amazon EC2, and IAM, enable them to support millions of customers-transacting millions of daily purchases-without a dedicated infrastructure team.

STP16: Securing Microservices on Amazon ECS at BuzzFeed Dec 01, 2018

Show notes

As BuzzFeed transitioned to a microservice platform build on Amazon ECS, it needed to secure a growing number of independent internal apps. The first solution was an open source OAuth proxy service deployed in front of each app, but this approach had a number of issues. In this talk, software engineer, Shraya Ramani, discusses BuzzFeed's experience and introduces SSO, a centralized authentication solution, which elegantly solved the problems faced by the company.

STP17: Lending at Affirm: Powered by Machine Learning and Amazon EMR Dec 01, 2018

Show notes

At Affirm, we lend billions of dollars to millions of customers using data-driven machine learning models that are trained on years of lending history. Feature extraction and validation in a risk management context presents unique challenges, so we have built a pipeline to eliminate training/serving skew in both development and production systems, and to run extensive regression testing on proposed PRs. Running Spark on Amazon EMR clusters provides an economic means to speed up this process. In this talk, we review our signal extraction and decisioning systems architecture and discuss how we use AWS to scale our compute.

STP18: Reddit's Serverless & Compute Infrastructure at Scale Dec 01, 2018

Show notes

Reddit is one of the world's most trafficked websites, with over 330 million monthly active users. Reddit built the vast majority of its infrastructure on AWS to support its increasing growth. As the world shifts to video-first, Reddit has redesigned its website and launched a new video platform. In this talk, we discuss design paradigms, tradeoffs made, and lessons learned from rearchitecting one of the internet's top sites. See how this infrastructure operates at scale and how the team leverages ETS, serverless, and compute to serve over one billion videos a month. Complete Title: AWS re:Invent 2018: 1B Video Views a Month: Reddit's Serverless & Compute Infrastructure at Scale (STP18)

STP19: Managing 1,000 Servers without a Driver's License Dec 01, 2018

Show notes

eSports streaming platform, Bebo, has thousands of servers in 15 AWS Regions and a fully dynamic automatically scaling stack with good uptime. Its servers are also managed by a teenager who can't yet get a driver's license. In this talk, listen as Bebo CTO Furqan Rydhan and DevOps engineer Johnny Dallas discuss how 16-year-old Dallas, who graduated from San Francisco's Drew School two years early, built Bebo's fully automated infrastructure in two months and finds the time to balance coding with studying for the SATs.

Our TOPPODCAST Picks

Follow Us

Stay Connected

Related Podcasts

Links

Stay Connected