From Scikit​-learn to TensorFlow Estimators - PyData Meetup

Published on: Tuesday, 9 January 2018

Speaker: Karthik MSwamy

Recent advances in deep learning has piqued the interest in machine learning and applications surrounding it. Software programmers with no background in mathematics can now use simple and efficient numerical tools to implement programs capable of learning from multi-modal data. Scikit-learn is one of the famous libraries adopted for developing classical ML applications while TensorFlow is a modern numerical computation tool capable of running a wide range of algorithms, on a wide variety of hardware. In this talk, we discuss concrete examples of how developers using Scikit-learn could seamlessly transition to TensorFlow. We compare and contrast TensorFlow’s Estimator API with Scikit-learn, and discuss applications ranging from text classification to image classification​,​
using both frameworks.​ We will use Jupyter notebooks to develop our application and explore some visualisation techniques to get a better understanding of the data.

Dr. Karthik Muthuswamy is a research​ scientist​ in the Innovation Center Network at SAP, where he works on making SAP service ticketing intelligent.

Karthik has designed and developed machine learning solutions for problems ranging from algorithms for guiding autonomous vehicles to understanding the semantic meaning of sentences in documents. He gives talks and conducts workshops on machine learning for the developer community to reduce the entry barriers to developing machine learning applications.

Connect with Dr Karthik over:
L​inkedIn: ​/in/karthikmswamy
T​witter​​: @krtk​

Event Page:

Produced by Engineers.SG
Recorded by:

Help us caption & translate this video!