Traveloka x DSSG: NLP & Personalised Recommendations (Part 1)

Published on: Sunday, 9 June 2019

Links to slides:

Abstract 1:
In this talk, Deb Goswami (Data Science Lead at Traveloka) will walk through challenges in enabling high quality NLP for low resource languages such as Bahasa. These challenges include using word2vec (Transfer learning + Vanilla), Data Quality, Issues, Various types of LSTM results, Transition into Char-level embeddings, or types of ML metrics to use, as well as challenges in production.

Speaker 1 Bio:

Dr. Deb Goswami is a seasoned Machine Learning practitioner, with over 10 years of research and production experience into enabling large-scale AI systems.
He is currently building out data science at Traveloka, Indonesia's leading multi-billion dollar travel unicorn, where his teams are responsible for delivering advanced machine learning capabilities in Natural Language Processing, Computer Vision, Speech and Personalisation.
Prior to this, he has worked on production AI solutions across a variety of industries ranging from Oil and Gas, Telecoms, E-commerce and Advertising.

Abstract 2:
*Dynamic Personalised Recommendations: Data and Algorithms*
In this talk, Yu Xuan will be sharing the Smart Adaptive Recommendations algorithm. It can be used to serve both item similarity recommendations and dynamic personalised recommendations. Despite this, it is very straightforward conceptually and is a useful baseline for personalised recommendations. He will cover the data required, the models used and some implementation details.

Speaker 2 Bio:
Yu Xuan has more than 5 years experience working in Data Science and Machine Learning Engineering. He is currently a Data Scientist in Traveloka where he builds ranking and recommendation systems for products across Traveloka businesses. Previously, he has worked in Lazada and GovTech, handling data science problems in various domain such as language, images and fraud.

Produced by Engineers.SG

Help us caption & translate this video!