DataKRK 1/2019

• 6:00 PM - 6:30 PM - Networking
• 6:30 PM - 7:15 PM - Cutting edge improvements in sentiment analysis
• 7:15 PM - 7:30 PM - Short break
• 7:30 PM - 8:15 PM - The Challenges of Productizing Natural Language Processing
• 8:15 PM - 9:00 PM - Q&A, Networking

* Snacks and drinks will be served through the whole event!

Speaker: Jamie Morningstar, Qualtrics

Abstract:

Natural Language Processing is hard work - and making a product that’s reliable, multi-purpose, accurate, and intelligible enough to be used by tens of thousands of self-service clients across millions of datasets? That’s the daunting challenge that the Qualtrics Text iQ team took on three years ago.

Come learn from our mistakes and successes in productizing a commercial text analysis tool. We’ll talk about some lucky breaks, good assumptions, and dramatic missteps we’ve made along the way. We’ll discuss product definition, build/buy analysis, data platform constraints, tuning challenges, algorithm choices, and more.

Bio:

Jamie is the engineering manager over all NLP and ML teams at Qualtrics. She is based in Utah, USA and has enjoyed 15 years of experience in software development, engineering management, and product management across several different cloud-based products. Jamie founded Qualtrics Text iQ three years ago and has grown the product to serve over 35k users analyzing millions of documents monthly.

Speaker: Felix Fang, Qualtrics

Abstract:

Sentiment analysis is a common NLP task in text analytics. The results are useful for many downstream tasks such as customer review analysis and stock price predictions. In this talk we will discuss how we improve sentiment analysis at Qualtrics using state-of-the-art techniques such as transfer learning with pretrained deep language model and question-answer pair training on user feedback data.

Bio:

Felix is an machine learning engineer and team lead of the ML/AI lab at Qualtrics. He works at the Seattle office and primarily focuses on building and productionizing various ML models for NLP tasks, such as sentiment analysis, aspect extraction, text summarization. Felix is an ML enthusiast and organizes events such as ML reading groups to review state-of-the-art work in the academia and industry.