Lerner?—?using RL agents for test case scheduling

The Netflix TechBlog

We thought that reinforcement learning would be a promising approach that could provide great flexibility in the training process. Likewise it has very low requirements on the initial amount of training data. The agent training library is written in Python and supports versions 2.7,

Stuff The Internet Says On Scalability For November 23rd, 2018

High Scalability

Waqas Dhillon : The goal of in-database machine learning is to bring popular machine learning algorithms and advanced analytical functions directly to the data, where it most commonly resides – either in a data warehouse or a data lake. Wake up!

Seer: leveraging big data to navigate the complexity of performance debugging in cloud microservices

The Morning Paper

Seer uses a lightweight RPC-level tracing system to collect request traces and aggregate them in a Cassandra database. A DNN model is trained to recognise patterns in space and time that lead to QoS violations.

Meta-learning neural Bloom filters

The Morning Paper

For example, a Google Bigtable database usually contains one Bloom Filter per SSTable file. In the case of a large database we think this is warranted, a database with 100K separate set membership data structures will bene?t Meta-learning neural bloom filters Rae et al.,

Stuff The Internet Says On Scalability For August 3rd, 2018

High Scalability

Paco Nathan : Frankly, I’d feel a lot more comfortable sending my kids off to school in a self-driving bus if the machine learning models hadn’t been trained solely by Google’s proprietary data.

Organisational Dysfunctions Mirrored as Architectural Complexity

Strategic Tech

It sucks all of the data out of the APIs and stores it in their own local database. Don’t hesitate to contact me if you’re looking for engineering leadership or domain-driven design expertise, consulting, or training.

Infinitely scalable machine learning with Amazon SageMaker

All Things Distributed

For example, training on more data means more accurate models. Last re:Invent, to make the problem of authoring, training, and hosting ML models easier, faster, and more reliable, we launched Amazon SageMaker. " To make things even more challenging, a system that can handle a single large training job is not nearly good enough if training jobs are slow or expensive. Machine learning models are usually trained tens or hundreds of times.

How enterprises can successfully scale Agile development


Tasktop unlocks lifecycle data from these application tool silos by automatically compiling modeled lifecycle activity data into a single database. After attending training, you will have the knowledge to become a VSM ambassador at your organization or agency.

A foundational strategy pattern for analysis: MECE

O'Reilly Software

It's not exhaustive: we left out testers, and other roles depending on your organization, such as release engineers, database administrators, project managers, and so on. Imagine you've been enlisted to create a recommendation to the CTO for a new database system to replace your legacy system.

Stuff The Internet Says On Scalability For July 27th, 2018

High Scalability

That means every employee, contractor, and the thousands of agents who work at the company show up to work—team meetings, training seminars, onboarding sessions—all inside a virtual reality campus. Lance Gutteridge : The reason why almost no one encrypts their databases is one of the dirty secrets of IT. Hey, it's HighScalability time: Startup opportunity? Space Garbage Collection service. 18,000+ known Near-Earth Objects. NASA ). Do you like this sort of Stuff?

Data Mining Problems in Retail

Highly Scalable

can be estimated by means of classification and regression models trained on historical data for customers who have received incentives in the past and those who did not. Propensity models are regression and classification models trained on customer data.

Retail 175

Consumer-driven Coupling: Patterns and Anti-patterns

Strategic Tech

At first, the API developers looked at all of the data in the monolith’s database and decided that because all the data was in the same database they could provide a single API endpoint which contained all the data the app needed so it only had to make a single call when a page was loaded or updated.

How architecture evolves into strategy

O'Reilly Software

Sharding and partitioning the database affords greater performance and distribution but creates complexity that is difficult to manage. Talent strategy: how you source and retain talent, how you train them.

Introduction to Wait Statistics

SQL Performance

One other resource for helping you understand waits is an online training course I recorded for Pluralsight – see here. Back in 2014, I started a series of blog posts here to talk about specific wait types and what they do and don’t mean.

USENIX LISA 2018: CFP Now Open

Brendan Gregg

Submit #lisa18 talk, training, panel, and demo proposals by May 24 [link]. Join us for 3 days in Nashville at LISA'18. Post by Brendan Gregg and Rikki Endsley. USENIX’s LISA conference is the premier event for topics in production system engineering.

Expanding the Cloud: Introducing the AWS Asia Pacific (Seoul) Region

All Things Distributed

Mirae Asset Global Investments improved its web service environment and reduced annual management costs by 50% by consolidating the management of all web services, including servers, network, database, and security.

Games 83

Serverless at re:Invent 2017

The Symphonia

Serverless Aurora promises to be the SQL Database to partner with Lambda that we’ve been hoping for. It’s still in early access, so no concrete news to report here, but a lot of people will be happy to be able to use a SQL database that has the operational properties of DynamoDB.

Improving Customer Service with Amazon Connect and Amazon Lex

All Things Distributed

For instance, Zillow trains and retrains 7.5 AWS Lambda functions made the corresponding calls to the database and scheduling software, making sure that the interaction happened quickly and at extremely low cost.

AWS Pop-up Loft 2.0: Returning to San Francisco on October 1st

All Things Distributed

You’ll have access to training including hands-on bootcamps and labs, and 1:1 sessions with AWS Solutions Architects. Topics include Introduction to AWS, Big Data, Compute & Networking, Architecture, Mobile & Gaming, Databases, Operations, Security, and more.

Games 62

The Amazon.com 2010 Shareholder Letter Focusses on Technology.

All Things Distributed

To do so, weve leaned heavily on the core principles from the distributed systems and database research communities and invented from there. The diversity of products demands that we employ modern regression techniques like trained random forests of decision trees to flexibly incorporate thousands of product attributes at rank time. a Fast and Scalable NoSQL Database Service Designed for Internet Scale Applications. Job Openings in AWS - Senior Leader in Database Services.

Our Once and Future Wisdom: Re-acquiring Lost Institutional Knowledge

The Agile Manager

It took a lot of challenging the "why" on those specific portions of the value stream before they understood how a simple shared database would eliminate lots of no-value-added inventory control steps. More modestly, we can simply license technology to replace major portions of legacy systems, and train or hire experts in that technology. Last month we looked at the loss of institutional memory and the reasons for it. This month, we look at our options for re-acquiring it.

Sector Seven Is Clear

The Agile Manager

Industrial IT encourages specialization: It is easier for HR to hire, staff, train and procurement to contract for people in specialist roles. It's easy to become proficient in a narrow field, such as a specific database or programming language. Many years ago, there was a television ad that showed an intruder being chased through a building by two security guards. The guards chase him from room to room, and ultimately down a long hallway.