Manage interdisciplinary team of 30+ people, including interaction designers, visual designers, PoC engineers, user researchers, producers. Our lab has led UX commercialization efforts on Samsung Smart TV (core UX + new services such as Sports, Music, TV Plus, Universal Guide), art mode UX for Samsung's The Frame lifestyle TV, UX for Family Hub 2.0 refrigerator.
Led projects on next generation products and services for Samsung, incorporating:
competitive analysis
product strategy and business opportunity
concept development and user scenarios
formative and evaluative user research
design and development of proof of concept prototypes
Managed both interdisciplinary and technical teams, including full-time employees, contractors, external partners and interns.
01/2002 - 02/2011
Eastman
Kodak Company
(Employee from 06/2004; Consultant from 01/2002 to 06/2004), San Jose,
CA.
Researcher, Photographic Science & Technology
Center, Kodak R&D (10/2004 - present): conduct research in areas related to digital
photography, focusing on user interface and user experience.
Project Leader, New Business Development, Kodak R&D
(01/2002 - 10/2004): identify new business opportunities for Kodak; size
market,
build
business case, assess technical capabilities and leverage, lead
ideations. Lead a cross-site, geographically distributed team.
09/1997 - 09/2001
Software Production Research Department,
Bell Labs/Lucent Technologies,
Naperville,
IL. Member of the Technical
Staff: conduct research in several areas including: information
visualization, next generation
services for wireless handheld devices, collaborative systems, and
user interfaces.
06/1990 - 09/1996
Project FLAME (Foreign Language Applications in
the Multimedia
Environment), The
University of
Michigan, Ann Arbor, MI. Technical, Design, and
Educational
Specialist: Design and develop interactive multimedia applications for
foreign language and multi-cultural learning and instruction; oversee
maintenance of multimedia hardware; review and integrate new
technology.
Summer 1990
Aerospace Engineering and Computer Science
& Engineering
Departments, The
University of Michigan,
Ann Arbor, MI. Project Leader: led team of six in designing
and
developing software for promoting math and science to young
adolescents, with focus on females and minorities.
06/87 - 08/89
InterUniversity Consortium for Educational
Computing (ICEC),
Pittsburgh, PA. Programmer and Technical Support: Programmed;
provided technical software support for faculty, staff, and students at
member schools; gave software demonstrations; maintained electronic
distribution lists; wrote articles and edited a bimonthly newsletter.
Teaching
Experience
Winter 1992
Toolbook Workshops. Led a series of
three three-hour
workshops on the introduction and use of Toolbook with Multimedia
Extensions.
June 1991
MediaText Workshops at NECC'91 (National
Educational Computing
Conference, Phoenix, AZ) -- coordinated over 20 workshops
with over
500 participants.
Research
Projects
Projects
at Bell Labs
handiMessenger
awareness-enhanced
universal communication for wireless
handheld devices (joint work with Audris Mockus)
handiMessenger is an awareness-enhanced universal
communication
service for use on a wireless Personal Digital Assistant (PDA). The
service is designed to address the key problem of initiating a
successful and timely communication in collaborative work by
transparently integrating messaging, awareness, and communication
services to be accessible by a mobile user in any situation. Through
the handiMessenger service, users can access their messages (e.g.,
email, instant messages, connectIcons), check the current availability
of their team members, view individual and group calendars, access
corporate documents, and use simple click-to-communicate links to
initiate various modes of communication between a variety of endpoints.
This can all be done from the user's handheld wireless PDA. The tight
integration of awareness information with messages, calendars, and
communication is designed to:
maximize the probability that the most appropriate
communication
mode and target is selected (email, instant message, phone number), and
increase the likelihood that the communication attempt will
be
successful.
Within handiMessenger, we are also exploring the integration of next
generation voice services, such as third party call control for SIP
phones.
InfoStill
an information
visualization environment to support
the end-to-end process of analyzing data using information
visualization (joint work with Ken Cox,
Audris Mockus, Graham Wills, and Lichan Hong)
A task analysis of infoVis tasks indicated that
infoVis users spend a significant amount of time working on important
tasks outside of data exploration -- tasks such as preparing data,
collecting and ranking results, and creating a presentation of results.
Unfortunately, infoVis research typically focuses only on data
exploration. In the InfoStill project, we are thus
working
towards supporting the end-to-end process of infoVis data analysis. Our
new infoVis infrastructure allows users to:
organize their analysis along lines of inquiry and across
different types of data,
explore their data using multiple infoVis views in a
linked-views
paradigm, and
easily compose annotated, interactive presentations of
results in
a standard Web format (i.e., LiveDocs).
This type of support is essential to making infoVis more accessible to
non-infoVis experts.
LiveDocs
web-based information
visualization documents that provide
guided exploration of data for naive infoVis users
(joint work with Audris Mockus and Todd Graves)
We refer to our web-based presentation of results generated by
InfoStill
as LiveDocs. LiveDocs incorporate the interactive
infoVis
applet views created with mmvz into the context of
a standard
web page. In addition, we
have enhanced LiveDocs with mmvz scripts. This allows us to place
scripts for view manipulation directly into standard web links or form
elements. This is particularly useful for linking view manipulations
right next to the relevant textual inquiry (e.g., so users can simply
click a link to make data selections that help "tell a particular story
about the data"). Adding this feature to LiveDocs provides a simple way
of guiding readers in exploring data results.
mmvz
java-based visualization
library
(joint work with Ken Cox, Graham Wills, Audris Mockus, and
Lichan Hong)
mmvs is a new, component-based infoVis library written in
Java.
The characterization of, and requirements for this library were
determined through a commonality analysis. Our mmvz components are the
key infoVis components used in InfoStill and LiveDocs.
Natural
Language Interface to InfoStill
(Joint work with Ken Cox,
Lalita Jagadeesan, David Mantilla
and Beki Grinter)
Current infoVis interfaces typically restrict user input to
the
traditional mouse and keyboard. Recent technologies are now enabling
new modes of interaction such as speech. In this project, users can
manipulate views and pose data analysis inquiries to InfoStill
using our patented natural language interface and response approach.
Our interface uses the power of Sisl (a
domain-specific
language for multi-modal services) and the built-in features of IBM's Via
Voice to exploit the features of natural language while
complementing the standard mouse interface. Sisl
provides
flexibility on both the way and the order in which commands are
specified.
Task
Analysis of Users Analyzing Data Using Information
Visualization
A task analysis was conducted to identify the types of tasks
users
do when analyzing data with an existing infoVis tool. The goal of the
task analysis was to identify ways in which
to better support users in the process of analyzing
data using
information visualization.
In the task analysis, we observed five infoVis experts using an
existing infoVis environment (EDV: the Exploratory Data Visualizer).
The analysis revealed 44 different tasks divided into seven categories
of high-level analysis tasks:
re-orient: re-orientation tasks (when
data analysis
requires more than one session), and
other: statistics-based tasks.
When asked to rate the importance of these high-level tasks, users gave
highest ratings to planning, exploration, preparation, and presentation
tasks, followed by slightly lower ratings for statistics (other),
overlay
and re-orientation tasks. Overall, users indicated that they typically
spend
only 25% on average, and 40% at most, of their time on data
exploration.
This means that they spend more than half of their analysis time on
other
tasks, all of which are typically unsupported by an infoVis system. The
results of this task analysis were used in the design of our InfoStill
system.
Previous
Projects and Research
MMVIS: MultiMedia Visual
Information
Seeking: An interactive information visualization approach to
analyzing temporal relationship trends in data such as video. MMVIS
integrates a temporal visual query language (TVQL) with an abstract
temporal visualization (TViz) of results.
Project FLAME (Foreign Language Applications in the
Multimedia
Environment): designed and developed end-user programming tools for
supporting four pedagogical models (Teacher's Partner, Learner's
Partner, Exploratory Mode, and Creative Mode) of using and integrating
technology into foreign language and multicultural classes.