I've worked on a variety of projects at AWS, largely focused on the needs of Application networking customers at ECS. I designed and led large parts of development for ECS Service Connect (https://docs.aws.amazon.com/AmazonECS/latest/developerguide/service-connect.html), a managed Service mesh offering with virtually zero setup time for customers. I presented this feature on an episode of Containers from the Couch (see Media) and have helped shape the roadmap of this feature to further meet ECS customer needs.I'm now working on adding extensions to ECS applications, making it easier to solve problems in Networking, Security and Storage for customers.

X-Ray: http://tinyurl.com/mse8uyf. X-Ray is a reading feature on Kindle devices and apps that lets you explore the "Bones of the Book." See all the passages across a book that mention relevant entities(ideas, fictional characters, historical figures, places, or topics of interest)Technologies: Java, AWS(SWF, S3, DynamoDB, SNS, SQS), Hadoop MapReduce, Ruby-on-rails, OpenNLP, StanfordCoreNLPHighlights:Kindle CoreNLP service: Designed and implemented a system to vend NLP artifacts for all Kindle books to external teams. The vended artifacts allow consumers to use high quality NLP data to build content features without having to deal with the complexities of setting up NLP processing pipelines or complicated internal book formats. X-Ray for Chinese: Designed a Tokens data model which worked with Latin and CJK languages where token breaks are less clear than in latin languages. Upgraded nearly 100k latin language books to use the new data model.without any major outages.

X-Ray: http://tinyurl.com/mse8uyf. X-Ray is a reading feature on Kindle devices and apps that lets you explore the "Bones of the Book." See all the passages across a book that mention relevant entities(ideas, fictional characters, historical figures, places, or topics of interest)Technologies: Java, AWS(SWF, S3, DynamoDB, SNS, SQS), Ruby-on-rails, OpenNLP, StanfordCoreNLPHighlights:- X-Ray terms: Shipped the algorithm to detect useful non-character entities in a book (eg. Places, Organizations, interesting noun phrases) which used a combination of shallow parsing, statistical filtering and external reference matching. - X-Ray entities service: Co-designed a service to vend X-Ray entities data to external teams. Also, built an internal website using ruby-on-rails to examine and easily edit X-Ray data for Kindle quality teams.

Amazon Kindle - Kindle Learning Machines