PyData Helsinki Meetup #5 : NLP & Feature Store (March 2021)

We are delighted to announce the 5th edition of PyData Helsinki meetups. This month's talks will be NLP and Feature Store. And again, due to the current situation, the meetup will be held online.

The event will be held on Google Meet and it will be recorded. Please make sure your mic is MUTED when you join the meeting.

Google Meet link: https://meet.google.com/nwq-zcbs-qrn

Agenda:

• 6:30pm - 6:35pm : Greetings & Introduction
• 6:35pm - 7:00pm: Faster intent training for chatbot using SBERT
• 7:00pm - 7:10pm: Q&A
• 7:10pm - 7:40pm: How Feature Stores Enable Operational ML
• 7:40pm - 7:50pm: Q&A

-------------------------------------------

Talk #1: Faster intent training for chatbot using SBERT, by Le Hoang Minh Anh (Mia Le)

In order to chatbots to understand human language with the highest accuracy possible, you will have to train it on 10 - 20 example phrases per intent. But what if you can combine a database of past conversations with a very powerful NLP model - SentenceTransformers - to automate the training process by generating 20 similar sentences from only one phrase? Even more interesting, you can customize it to your language of choice: Finnish, Swedish, Vietnamese, etc. If you are curious to know how it can be done, please kindly join the talk and I'll tell you my little secret.

Mia Le is a Machine Learning Consultant for a chatbot startup called Hello Ebbot based in Stockholm and also a Junior AI Scientist Intern at Panda Training. she started researching about NLP and chatbots for her Bachelor's thesis in December 2019 and decided to follow this career path. She is extremely passionate about NLP and sharing my achievement with the community.

Talk #2: How Feature Stores Enable Operational ML, by David Hershey

In this presentation, David will outline how a feature store can help deal with common challenges data scientists are faced with when operationalizing ML Models. He will expand upon the differences between Operational ML vs Analytical ML.

David David Hershey is a solutions architect at Tecton, where he helps companies integrate Tecton's feature store into their ML stack. He has spent the last two years in the MLOps space, working on building ML platforms and helping teams build out their ML capabilities. David has an MS in Machine Learning from Stanford, and before he came to Tecton he worked at Ford Motor Company as a machine learning engineer.

-------------------------------------------

PyData is a community for developers and users of open-source data tools. PyData is an educational program of NumFOCUS, a 501(c)3 non-profit organization in the United States. The PyData Code of Conduct governs this meetup.