List of videos

This video was recorded at Code BEAM America 2021 - https://codesync.global/conferences/code-beam-sf-2021/ Survivor bias in software testing, observability & monitoring | Jan Cieśla - MongooseIM Developer at Erlang Solutios ABSTRACT Drop out from college to become a millionaire! This is an example of a Survivor Bias which is a logical error in our understanding of the world, the assumption that success tells us the whole story about the process. Real world examples help us see how prone we are to it. During this presentation I want to talk about it with respect to testing, observability and monitoring. We will learn the techniques on dealing with it to make tools that are more reliable and exhibit our internal biases to lower extent. OBJECTIVES: Bring the phenomenon of survivor bias closer to the audience. Help realise how it can impact our judgement and what we can do about it. AUDIENCE: Developers of all shapes and sizes. • Timecodes 00:00 - 02:50 - Intro 02:51 - 10:55 - Statistical Research Group 10:56 - 13:29 - Survivorship Bias 13:30 - 14:12 - What to do? 14:13 - 15:25 - Products 15:26 - 16:58 - Testing 16:59 - 18:45 - Monitoring 18:46 - 23:18 - Observability 23:19 - 24:40 - QnA • Follow us on social: Twitter: https://twitter.com/CodeBEAMio LinkedIn: https://www.linkedin.com/company/27159258 • Looking for a unique learning experience? Attend the next Code Sync conference near you! See what's coming up at: https://codesync.global • SUBSCRIBE TO OUR CHANNEL https://www.youtube.com/channel/UC47eUBNO8KBH_V8AfowOWOw See what's coming up at: https://codesync.global

This video was recorded at Code BEAM America 2021 - https://codesync.global/conferences/code-beam-sf-2021/ Building a Serverless platform for Edge Computing | Benoit Chesneau - owner of Enki Multimedia ABSTRACT After the feedback from our customers at Enki Multimedia, I started in Erlang a new opensource "On-Premise Serverless" platform with an emphasis on moving the data and its application, next to the consumer in a P2P manner. Based on barrel it allows you to build a P2P messaging and data platform using replication and synchronisation mechanisms. Eg. a multi-homing application, a decentralised message hub, a signalling platform, … . I will present the platform and how to extend it in Erlang but not only. OBJECTIVE I will present some key mechanisms of the platform and how it can be extended in Erlang but not only. At the end the audience should understand how Erlang is useful in such platform and how to use and customise the platform to fit their needs. AUDIENCE Anyone interested in edge commuting and serverless platforms • Timecodes 00:00 - 02:16 - Intro 02:17 - 02:51 - What is Edge Computing? 02:52 - 03:26 - What is Serverless Computing? 03:27 - 07:26 - Case Studies 07:27 - 11:15 - Keys 11:16 - 13:35 - The DP Engine 13:36 - 15:42 - The Application Nodes 15:43 - 19:10 - Persist 19:11 - 20:52 - The Assembly Engine 20:53 - 22:42 - Key Components 22:43 - 24:09 - QnA • Follow us on social: Twitter: https://twitter.com/CodeBEAMio LinkedIn: https://www.linkedin.com/company/27159258 • Looking for a unique learning experience? Attend the next Code Sync conference near you! See what's coming up at: https://codesync.global • SUBSCRIBE TO OUR CHANNEL https://www.youtube.com/channel/UC47eUBNO8KBH_V8AfowOWOw See what's coming up at: https://codesync.global

This video was recorded at Code BEAM America 2021 - https://codesync.global/conferences/code-beam-sf-2021/ When you have to make it faster | Marc Sugiyama & Niko Tsikoudis - Experienced Erlang engineer & Software Engineer at Datometry ABSTRACT Our product, Hyper-Q, enables modern cloud data warehouses to emulate Teradata, aiding with the migration of on-premises solutions to the cloud. Hyper-Q translates existing applications' queries in real-time to run natively in the cloud and converts data received from the cloud into the format needed by the application. We recently missed service level agreements with a customer because our product, Hyper-Q, added significant overhead to a data warehouse application. Through careful investigation, using profilers, continuous improvement, and focused engineering considering Erlang characteristics, we doubled our performance and improved system scalability. This talk describes the steps and techniques we used to understand where to focus our engineering efforts and beat our service level agreements. • Timecodes 00:00 - 04:44 - Introduction to Datometry/ Hyper Q 04:45 - 06:05 - Customer Incident 06:06 - 08:25 - In House Reproduction 08:26 - 09:55 - Identifying Dominant Cost 09:56 - 15:22 - How do we make Data Conversion Faster? 15:23 - 17:21 - Optimization Example 17:22 - 17:52 - Result 17:53 - 19:30 - Profiling Takeaways 19:31 - 23:50 - Binary Optimizations 23:51 - 26:22 - Improvements 26:23 - 27:16 - Optimizations 27:17 - 28:06 - Next Steps 28:07 - 35:14 - QnA • Follow us on social: Twitter: https://twitter.com/CodeBEAMio LinkedIn: https://www.linkedin.com/company/27159258 • Looking for a unique learning experience? Attend the next Code Sync conference near you! See what's coming up at: https://codesync.global • SUBSCRIBE TO OUR CHANNEL https://www.youtube.com/channel/UC47eUBNO8KBH_V8AfowOWOw See what's coming up at: https://codesync.global

This video was recorded at Code BEAM America 2021 - https://codesync.global/conferences/code-beam-sf-2021/ Typed_erlc: Prototype of safe & fast compiler for Erlang | Dmytro Lytovchenko - Senior Developer at Erlang Solutions ABSTRACT Presenting a new compiler, which consumes classic Erlang syntax and outputs standard BEAM files. Explaining reasoning behind the project and what problems it is designed to solve, and what problems it may potentially solve in the future. Comparing to other compilers and other languages on the BEAM. This opens possibilities for new safety and style checks, new optimizations, new code generation backends (did i hear someone say WASM?), etc. OBJECTIVES Show the project in the making. Explain why it exists and how it is good for an average presentation listener. Explain how it will plug in into their tool chain and will help improve their code quality. AUDIENCE Erlang developers, mostly working/maintaining heaps of legacy code. • Timecodes 00:00 - 01:43 - Introduction 01:44 - 06:55 - I have a Dream 06:55 - 09:13 - What is this Project? 09:14 - 11:34 - Why it Exists? 11:35 - 14:27 - Progress 14:28 - 16:25 - H-M Type Inference 16:26 - 18:13 - Bidirectional Type Inference 18:14 - 20:23 - Distinctive Features 20:23 - 24:32 - Future Ideas 24:33 - 25:54 - When? 23:53 - 27:06 - Where? 27:06 - 39:11 - QnA • Follow us on social: Twitter: https://twitter.com/CodeBEAMio LinkedIn: https://www.linkedin.com/company/27159258 • Looking for a unique learning experience? Attend the next Code Sync conference near you! See what's coming up at: https://codesync.global • SUBSCRIBE TO OUR CHANNEL https://www.youtube.com/channel/UC47eUBNO8KBH_V8AfowOWOw See what's coming up at: https://codesync.global

This video was recorded at Code BEAM America 2021 - https://codesync.global/conferences/code-beam-sf-2021/ WebRTC redefined - a new SFU API for Elixir from Membrane Framework | Michał Śledź - Author of WebRTC Implementation in Elixir ABSTRACT Creating video conferencing systems has always been challenging. While the most used standard - WebRTC requires from the user quite a deep understanding of what is going on under the hood, the new abstraction layer created in Membrane Framework allows us to hide all the low level, network connected things. During this talk, I will go through the new SFU API and how it makes the developer experience much easier, both when creating a new Phoenix application and integrating with existing systems. OBJECTIVES: Presentation of the new SFU API and its usage. AUDIENCE: Everyone who is interested in designing APIs, abstraction layers or is willing to create or integrate a video conferencing platform with its existing system. • Timecodes 00:00 - 00:57 - Table of Contents 00:58 - 01:18 - What is Membrane_Framework 01:19 - 03:18 - Membrance Architecture 03:19 - 04:51 - What is WebRTC 04:52 - 06:44 - P2P Scenario 06:45 - 07:45 - WebRTC Protocol Stack 07:46 - 17:11 - SDP Example 17:12 - 17:31 - Problems of WebRTC 17:32 - 18:16 - What is Membrance RTC Engine? 18:17 - 19:59 - Architecture 20:00 - 20:45 - Client Side API 20:45 - 21:48 - Callbacks 21:49 - 25:50 - Functions 25:51 - 27:16 - What does Membrane RTC Engine gives you? 27:17 - 33:21 - One more Abstract Layer 33:22 - 34:03 - Links • Follow us on social: Twitter: https://twitter.com/CodeBEAMio LinkedIn: https://www.linkedin.com/company/27159258 • Looking for a unique learning experience? Attend the next Code Sync conference near you! See what's coming up at: https://codesync.global • SUBSCRIBE TO OUR CHANNEL https://www.youtube.com/channel/UC47eUBNO8KBH_V8AfowOWOw See what's coming up at: https://codesync.global

This video was recorded at Code BEAM America 2021 - https://codesync.global/conferences/code-beam-sf-2021/ Ditch the DB Using In Memory Replicated Distributed State| William Huba - CoFounder of Jumpwire ABSTRACT When data needs to be accessible from every node on a cluster, a DB is often a necessary choice. But what if you want to avoid that extra dependency and keep everything in-memory? The library Horde utilizes a technique called delta CRDTs for distributed supervisors and registries. In this talk you’ll learn how delta CRDTs work (using Horde as an example), how to leverage delta CRDTs for your own synchronized state, and the tradeoffs for doing so versus a traditional database. OBJECTIVE: Explain the concept of Delta State Replicated Data Types Detail the tradeoffs between using the DeltraCrdt library and an external database Teach how to use delta CRDTs as a building block for application state, using Horde and other custom code examples for illustration AUDIENCE: Experienced software engineers who have at least some Elixir experience and work with (or are interested in) the distributed properties of the BEAM. • Timecodes 00:00 - 03:09 - Intro 03:08 - 08:41 - Different Types of CRDTs 08:42 - 10:05 - Use Cases 10:06 - 11:57 - DeltraCrdt (The Library) 11:58 - 13:40 - Horde 13:42 - 15:05 - Merkle Tree 15:06 - 24:00 - Use Cases - Specifically 24:01 - 26:52 - Supervision Tree 26:53 - 29:52 - Future Work 29:53 - 34:03 - QnA • Follow us on social: Twitter: https://twitter.com/CodeBEAMio LinkedIn: https://www.linkedin.com/company/27159258 • Looking for a unique learning experience? Attend the next Code Sync conference near you! See what's coming up at: https://codesync.global • SUBSCRIBE TO OUR CHANNEL https://www.youtube.com/channel/UC47eUBNO8KBH_V8AfowOWOw See what's coming up at: https://codesync.global

This video was recorded at Code BEAM America 2021 - https://codesync.global/conferences/code-beam-sf-2021/ Practical uses for Metaprogramming in a highly trafficked web API | Jacqueline Manzi - Engineering Manager at MUX ABSTRACT Metaprogramming can often be more confusing than it is beneficial. In this talk, we will explore the basics of metaprogramming in Elixir and how it can be used in practical ways within layers of a highly trafficked web API. OBJECTIVES: Introduction to metaprogramming basics in Elixir Pros and Cons Examples of practical use cases in layers of a highly trafficked API AUDIENCE: Developers that want to know more about metaprogramming in Elixir and various use cases that can be applied to web API layers. • Timecodes 00:00 - 01:05 - Intro 01:06 - 05:09 - What is Metaprogramming 05:10 - 07:39 - Writing in Macros 07:40 - 09:17 - Authorization 09:18 - 13:37 - Phoenix Controller 13:38 - 15:39 - Pipeline your Router 15:40 - 18:08 - Consolidation with Macros 18:09 - 23:06 - Validation 23:07 - 24:59 - QnA • Follow us on social: Twitter: https://twitter.com/CodeBEAMio LinkedIn: https://www.linkedin.com/company/27159258 • Looking for a unique learning experience? Attend the next Code Sync conference near you! See what's coming up at: https://codesync.global • SUBSCRIBE TO OUR CHANNEL https://www.youtube.com/channel/UC47eUBNO8KBH_V8AfowOWOw See what's coming up at: https://codesync.global

The conference opening by chairwoman Christine Winter.

According to the 2021 State of AI report, "transformers have emerged as a general-purpose architecture for ML. Not just for Natural Language Processing, but also Speech, Computer Vision or even protein structure prediction." Indeed, the Transformer architecture has proven very efficient on a wide variety of Machine Learning tasks. But how can we keep up with the frantic pace of innovation? Do we really need expert skills to leverage these state-of-the-art models? Or is there a shorter path to creating business value in less time? In this code-level talk, we'll show you how to quickly build and deploy machine learning applications based on state-of-the-art Transformers models. Along the way, you'll learn about the portfolio of open source and commercial Hugging Face solutions, and how they can help you deliver high-quality machine learning solutions faster than ever before.