Can Liu

Generative recommendation system

- Contributed to improving RAG in Amazon's foundational model Nova. SFT/ Reward modeling. - Launched the News Q&A experience in Alexa, focusing on fresh and emerging events. Led the science of news event identification. - Contributed to improving NLU in Alexa Skills Kit (intent classification, named entity recognition).

Amazon Echo spoken dialog system. Research in Natural Language Understanding: semantic parsing

Amazon Echo spoken dialog system. Research in Natural Language Understanding: distant supervision

Adaptive Language Learning. Designed and implemented an adaptive learning algorithm, which models leaner’s language skills, and automatically adapt the curriculum. The prototype was realized as a web application.

Area: Natural Language Processing applied in Concept Networking •Responsibility: Work with research scientists in iLabs-LexisNexis. Implement research ideas and build.product prototype. Project involved: Word classification, Legal Concept Networking. Responsibility includes: parsing XML, building corpus, process data, visualizing and analyzing result. •LexisNexis has the world’s largest electronic database for legal and public-records related information. It provides computer-assisted services for legal researchers, attorneys and law school students.

Implemented a ‘Internet of Things’ project that controls the public facilities and hardware in Sichuan University.

I obtained candidacy in 2014 and moved to industry in 2016, because the opportunity to build the first voice-based assistant Alexa was too exciting.

Machine learning/ NLP