JAWS FESTA 東海道 2016 で講演させて頂いた資料です。Datadogは今年の7月に、お客様のシステムをモニタリングしてきた経験を踏まえてMonitoring in the CloudというホワイトペーパーをAWSのサポートの元、公開しました。本セッションでは、そのホワイトペーパーをベースにクラウドインフラのモニタリング理論を日本語で解説します。
Real world machine learning with Java for Fumankaitori.comMathieu Dumoulin
This document summarizes a presentation about using machine learning in Java 8 at Fumankaitori.com. The presentation introduces the speaker and their company, which collects user dissatisfaction posts and rewards users with points that can be exchanged for coupons. Their goal was to automate point assignment for posts using machine learning instead of manual rules. They trained an XGBoost model in DataRobot that achieved their goal of predicting points within 5 of human labels. For production, they achieved similar performance using H2O to train a gradient boosted machine model and generate a prediction POJO for low latency predictions. The presentation emphasizes that machine learning is possible for any Java engineer and that Java 8 features like streams make it a good choice for real
Kotlin is a statically typed programming language that runs on the JVM and is fully interoperable with Java. The document discusses some key reasons why Java engineers should consider Kotlin, including that it is more concise, safe, and versatile compared to Java. It also provides examples showing how Kotlin code can be more concise than equivalent Java code through features like data classes, default values, named arguments, and standard library functions.
This document discusses several popular Java libraries including:
- Dependency injection frameworks like Guice and Spring
- Logging with SLF4J
- Collections and utilities with Guava
- HTTP clients like OkHttp
- Reactive programming with RxJava
- REST with Retrofit
- Asynchronous programming with JDeferred
- Event handling with MBassador
- Code generation with Lombok and ByteBuddy
- Testing utilities like JUnitParams, Mockito, Jukito, and Spock
- Waiting assertions with Awaitility and REST testing with Rest-assured.