亚麻Alexa选组求问

拿到了亚麻Alexa组的offer, HR现提供了以下几个组供选择。
本人小白，还请各位懂的童鞋给建议~
本人三年工作经验。比较希望去一个能多学点东西，oncall不那么多的组。

Knowledge Extraction and Understanding (KEANU)

Provides horizontal services at scale to improve the user experience by understanding what users want: from the questions they ask, the answers we give and the wider context.  We identify and understand different content sources and orchestrate across them to provide the best response.  This is enabled by a combination of machine learning, natural language processing and semantic algorithms that extract knowledge to determine the intent of users, retrieve relevant content (in a variety of forms), rank answers and provide automated knowledge acquisition to enrich our knowledge base using  both structured and unstructured data sources.

·         Understanding: Understand customer utterances semantically - Enables the entire Q&A pipeline in different languages.  From a short utterance, we extract relevant entities, determine user intent and translate natural language into structured queries so a KB or system can provide answers.

·         Structured Extraction: Increasing the capability of our KB to answer  - Increases the capability of our knowledgebase to answer questions about items that users are interested in through automated structured and semi-structured data extraction. Improving recency, consistency and demand weighted entity completeness of the knowledgebase.

·         Search-Based Question Answering (SBQA): Searching, retrieving - We provide new methods for retrieving content, extracting and ranking answers to give users a response when a question is not understood semantically, or the knowledgebase (or other systems) cannot provide an answer.

·         FOUNT: Extracting structure where there is none - Provides the bridge from unstructured content to actionable knowledge by extracting entities, relations and facts with high precision.

·         Orchestration: Intelligently sourcing the answer - Coordinating between different methods of understanding and utterance and intent to channel the right answer.

Knowledge Graph and NLG

We own the Alexa Knowledge Graph (of facts, entities and relations) and we're extending our Natural Language Generation system, which transforms semantic representations of what the system to say into fluent natural language output.

·         Knowledge Graph –We run the Alexa Knowledge Graph (of facts, entities and relations) at Alexa scale. We provide the query API over the Knowledge Graph and the query language reference specification. We are working on enabling customers to extend the Knowledge Graph with their own data.

·         Natural Language Generation (NLG) –We focus on providing human understandable answers to questions. We strive to provide a best in class NLG system that generates accurate, fluent, engaging natural language from machine readable data on behalf of a wide range of customers in an increasingly self-service manner.

·         Answer –We focus on providing the information needed to answer semantic questions within the Alexa Info domain. We augment our answers with relevant information that enhances the customer experience.

·         Knowledge Curation Environments – We own the ability for the Alexa Knowledge Graph to support curation environments to enable users of the knowledge graph to run experiments with candidate knowledge before their changes are applied to the production database.

Knowledge Workflows

Our mission is to drive down the skill required to improve Alexa's question-answering ability, towards a north star of zero-touch. We do this by creating simplified interfaces for Alexa Information knowledge engineers, other Amazon teams, AMDS and members of the public. Behind the interfaces we automate where possible, for example replacing AMDS workflows with classifiers.

·         Modelling Infrastructure – We build and source tools to make model debugging, training, building and testing fast and simple for all ML and rules-based systems used within Alexa Knowledge.

·         Getafix and Admin Site –  We build workflows for non-specialist workers to improve Alexa’s question-answering ability and for Amazon teams to debug and manually curate data issues. We find ways to streamline the workflows with automation and ML.

Knowledge Base

We focus on providing a highly scalable self-service entry point for semantic knowledge into Alexa that ensures its consistency and availability for cross domain Q&A experiences and the retrieval of entities.

·         Consistency –We ensure a semantically consistent and complete Knowledge Base for Q&A. We provide an extensible platform for defining consistency for our users.

·         Access –We enable developers to bring domain knowledge into Alexa in familiar ways in multiple languages promoting reuse of existing data, so that other teams have an ever increasing amount of knowledge to work with at an ever decreasing cost of development

为啥说Alexa 没什么oncall？ Alexa有oncall 啊～而且有的组一点也不轻

最初的梦想2013 发表于 2018-6-13 22:14
有人知道Alexa的machine learning team怎么样？刚拿到offer，听起来挺牛的样子。

请问楼主面了吗，我最近也收到了面试，请问楼主有什么经验可以分享吗，感谢~

lindaxiong 发表于 2018-3-17 11:09
请问有人知道关于alex Knowledge engineer 的信息吗？
马上要onsite
非常感谢！！

请问你后来有拿到offer吗？我最近刚收到电面，可以求问一下经验吗？

自己顶一下。求推荐~

Alexa没什么oncall吧，lz是nlp方向的么？还是sde？

吃洋葱吧 发表于 2017-12-12 06:26
Alexa没什么oncall吧，lz是nlp方向的么？还是sde？

我是SDE . Java developer

清风疏桐 发表于 2017-12-12 06:27
我是SDE . Java developer

可以加我一下微信吗？想问一下alexa的情况，我的微信Leo07591553151，感谢楼主！！！

请问有人知道关于alex Knowledge engineer 的信息吗？
马上要onsite
非常感谢！！

有人知道Alexa的machine learning team怎么样？刚拿到offer，听起来挺牛的样子。

你有解答了吗？我知道这几个组的信息