Free

[FREE] ODSC East 2019 Conference Meetup

Event Information

Share this event

Date and Time

Location

Location

Hynes Convention Center

900 Boylston Street

Boston, MA 02115

View Map

Event description

Description

We are excited to invite you to join together with your fellow data scientists on May 1st at ODSC East Conference Meetup - Supported By Zepl, as we feature three 40 minutes talks presented by:

  • Moon soo Lee, CTO at Zepl

  • Jess Stauth, Research Managing Director at Quantopian

  • Heming Zhen, Data Engineer at Insight Data Science


Schedule:

6:00 pm - 6:30 pm - Networking
6:30 pm - 7:10 pm - Session 1
7:15 pm - 7:55 pm - Session 2
8:00 pm - 8:40 pm - Session 3


Full Agenda:

Session 1 - Productionizing Data Science Models Using Zeppelin Notebooks - 40 mins

Speaker: Moon soo Lee, CTO at Zepl - https://www.linkedin.com/in/moonsoo-lee-4982a511/

Bio:

Abstract:
TBD


Session 2 - Data Modeling the Stock Market Today - Common Pitfalls to Avoid - 40 mins

Speaker:
Jess Stauth, Research Managing Director at Quantopian - https://www.linkedin.com/in/jessicastauth/

Bio:
Dr. Jessica Stauth is the Managing Director of Portfolio Management and Research at Quantopian, a crowd-sourced quantitative investment firm, that inspires talented people from around the world to write investment algorithms. Jess and her team are in charge of selecting the algorithms from the Quantopian community, for our portfolio. Quantopian offers license agreements for algorithms that fit our investment strategy, and the licensing authors are paid based on their strategy's individual performance. Previously she has worked as an equity quant analyst at the StarMine Corporation and as a Director of Quant Product Strategy for Thomson Reuters prior to joining Quantopian in August of 2013. Jess holds a Ph.D. from UC Berkeley in Biophysics.

Abstract:
The lure of creating models to predict the stock market has drawn talent from fields beyond finance and economics, reaching into disciplines such as physics, computational chemistry, applied mathematics, electrical engineering and perhaps most recent statistics and what we now refer to as data science. The attraction is clear - the stock market (and the economy/internet at large) throws off massive and ever-increasing reams of data from garden variety time-series to complex structured data sets like quarterly financials, to unstructured data sets like conference call transcripts, news articles and of course — tweets! While all this data holds promise - it also holds traps and blind alleys that can be tricky to avoid. In this session we’ll review some of the common (but not easy!) pitfalls to avoid in creating models for predicting stock returns; overfitting & exploding model complexity, non-stationary processes, time-travel illusions, under-estimation of real-world costs, and as many more as we have time to cover.


Session 3 - Deploying deep learning model at scale with Docker and Kubernetes - 40 mins

Speaker:
Heming Zhen, Data Engineer at Insight Data Science - https://www.linkedin.com/in/hemingzhen

Bio:
Dr. Heming Zhen earned his bachelor’s degree in physics from University of Science and Technology of China before he came to the United States to pursue his PhD in Medical Physics at University of Wisconsin - Madison. After graduation Heming served as an assistant professor of medical physics at Rush University, then moved on to become an assistant professor of medical physics at Boston University Medical Center. Heming has conducted various research projects, including quality assurance in radiation therapy, cancer survival analysis, and machine learning in radiation therapy treatment planning. Following his interest in the field of data, Heming recently worked at Insight Data Science as a data engineering fellow. After completing the fellowship program, he joined Insight Data Science as the data engineering program director in Boston. Heming is very interested in the latest Machine Learning/Deep Learning models, as well as engineering tools and techniques to deploy these models at scale.

Abstract:
In this presentation, I will demonstrate a project I built in 4 weeks as a Data Engineering Fellow at Insight Data Science. In this project, I trained a Keras image classification model and deployed it at scale on AWS. This talk consists of two parts. In part I, I will give a brief introduction to several key concepts and tools that are essential to the project. The first concept is using transfer learning and image augmentation to train model with a limited dataset. Secondly, I will highlight using TensorFlow Serving to serve a pre-trained model as an API. Additionally, I will review the use of container technology to distribute application while removing dependency issues. Finally and most importantly, I will discuss how to utilize Kubernetes to scale up container deployment with fault-tolerance. In Part II, I will demonstrate details on how I trained and deployed the model, the engineering challenges I faced when trying to scale up the system throughput, and how I used the aforementioned tools to overcome these challenges.

This event is Supported by Zepl

Date and Time

Location

Hynes Convention Center

900 Boylston Street

Boston, MA 02115

View Map

Save This Event

Event Saved