2016.04.18
National Taiwan University
of Science and Technology
15:30 - 17:00
NTUST IB-202
Speaker : Dr. Heng Chu
Vice President, Architecture
Strategic Investment Products and Data
Fidelity Investments
Bio:
Dr. Heng Chu is a VP of Architecture managing a team of architects
responsible for the analytics platform architecture and strategy for the
Strategic Investment Products and Data group in Fidelity Investments.
The platform covers analytical applications and data including Big
Data, batch and real-time advanced analytics, and an omni-channel
personalization system.
Prior to his current role, Dr. Heng Chu was the Chief Architect for the
Customer Data Analytical Environment in Fidelity Investments. He was
responsible for designing the architecture and delivering a Big Data
analytical environment. This environment consists of MPP data warehouse
and Hadoop cluster that provide integrated data and processes
for Big Data descriptive and predictive analytics on structured and
unstructured customer data.
Before joining Fidelity in 2012, Dr. Chu was a Senior Technical Staff
Member at IBM with 15 years of experience where he served in
various leadership roles in software strategy, solution architecture, and
development in Data, Cloud, and Middleware areas. Heng has extensive
experience in leading development of multi-tier software application,
and managing cross-organization efforts including a technical
review of IBM Cloud offerings in 2010. Prior to IBM, Dr. Chu spent 3
years in software startups leading software solution development.
Heng Chu holds 7 IT patents, published several peer-reviewed papers,
and received corporate awards from IBM and Fidelity including IBM
Outstanding Technical Achievement Award, IBM Outstanding Innovation
Award, and Fidelity Impact Player Award. Dr. Chu earned his Ph.D.
degree in Computer Science (in the field of automated reasoning in
Artificial Intelligence) from the University of North Carolina at Chapel
Hill. He also has been a Project Management Profession (PMP) since
2001.
Presentation Abstract:
This presentation describes the architecture design of a Big Data
platform for descriptive and predictive analytics for omni-channel
customer personalization. We will show in details how we started with
a Data Warehouse, and evolved into a Big Data platform that consists
of Data Warehouse, MPP database, Hadoop ecosystem, data streaming,
and NoSQL. With the initial capabilities of descriptive analytics and
small predictive analytics, the platform has grown to handle Big Data
with integrated architecture and processes for descriptive analytics
and predictive modeling lifecycle. Through mixed batch and low-latency
analytics, it has enabled us to better understand and serve our
customers through multi-channel personalization.
We had challenges in building this platform – for example, integration
of various data components, meeting business needs in mixed batch
and low-latency analytics, and growing the platform quickly in a
cost-effective fashion. We will share the key architecture design
patterns that addressed those challenges.
Big Data adoption needs to consider existing data environment,
processes, and skill sets. Our experience shows a successful adoption
that should be both revolution and evolution, and it is not an easy
task. We will share technical and business lessons learned during our
Big Data journey. After this session, you should be able to apply the
architecture patterns in your Big Data projects, and deliver business
values quickly.
