Portfolio projects

The following is a selected group of portfolio projects that I am particularly fond of; a lion's share of which I was fortunate enough to be involved in for complete life cycles, from inception to leading the project to fruition, and catapulting it into markets. In most projects, I was deeply involved in technical design and development, as well. Please see the credits section.

Securio

Securio -- Cloud and Entrprise network security
Securio uses Machine Learning to prohibits information leakage and data theft from computer networks and the cloud.

FPGA based deep packet inspection

GV 2020 deep packet inspection using FPGA
The GV 2020 is a Deep Packet Inspection (DPI) appliance for computer networks. It can scan and apply security policies at network line rates of up to 10 Gbps.

Comversa eCommerce platform

CEPM eCommerce platform
Comversa, a B2B eCommerce platform. Integrated into corporate ERP of both trading partners and Ariba. Services are offered through SOA architecture. Backend optimization algorithms minimizes packaging dimensions and allow business to maximize freight profitability.

Error Management System (EMS)

Error Management System
Reduce revenue leaking and labor required to resolve processing errors for organizations with very high transaction volume. Machine Learning observes users corrective actions and applies these automatically on future transactions.

Telephony data analytics

Telephony Data Analytics
Data analytics package geared towards high volume data streams and telephony related business questions.

Get the Music you want

Get the music you want product conceptual design
Consumer electronics device that acquires audio snippets and connects listeners to the original song. Deep Learning is used to optimize matching accuracy over time.

Thinking Cashier

Thinking Cashier
Thinking Cashier is a computer expansion card that transforms a regular PC to a smart Point of Sale.

Up2Date

Up2Date Real-world clock for slot-less computers
Real-world clock for slot-less computers.

HelpU

QQQ
Online tutorial and OS wrapper to operate character based computers for novice users.

Packman AI

Pacman AI agent
Artificial Intelligence (AI) Packman agent. The algorithm participated in nightly games against other agents (all against all and finals between Stanford and WUSTL) and was never defeated. It continued to play during two additional semesters (with no change to the algorithm) while others designed to defeat it. It was beaten once.

VLSI 8-bit RISC CPU

VLSI 8-bit RISC CPU
VLSI design and development of RISC CPU using AMI 0.6 micron technology. Fits within 1.5 mm square silicon area.

Visual Office

Visual Office
Visual Office (VO) is an all-encompassing ERP solution for distributed organizations such as chain stores. VO provides visibility and functionality for back-office operations, order management, accounting, inventory management, and consumer analytics. It is an enterprise-wide system that builds knowledge from distributed data objects and supports multi-site communication with data coherency. VO included a UI compiler that builds multilingual interfaces to the system.

Science Nature And Photography (SNAP)

Science Nature And Photography (SNAP)
We (Joy Chiang, Bryn Beez, Adam Kraft, and yours truly) completed the development of SNAP in April 2010, an iPhone app that enables citizen-scientists to send images to AMOS. SNAP tags the image with GPS location, reference to Earth magnetic pole, as well as other annotated data and uploads it to AMOS, expanding its reach and enabling new research. Scientists desiring images from certain locations can specify their requests and specific instructions using a web-app. iPhone users subscribe to these feeds based on their travel location and interests. The downside is that we did not think to turn this into a general purpose app. Bravo to Instagram leadership, who launched their app in Oct. 2010.

Plant

Plant system
Shtil (plant in Hebrew) is a system for management of agriculture businesses. It facilitates client management, invoicing, and annual planting and seed-breeding success analysis.

Next Release

Our software will make coffee in the next release
My software will make coffee in the next release.

SOA Integration legacy with modern technologies

QQQ
Corporate Express legacy systems were abstracted as providing functionalities accessible via Service Oriented Architecture (BUS of services). New systems were built to offer a discrete set of services available through the BUS. This architecture yielded tremendous flexibility and reuse of services. Development cycles shortened and quality increased.

Probe

Probe system
Probe is a system for help-desk, service dispatching, and technicians scheduling. Its greatest value was in its free text searchable knowledge base.

Ay

Ay medical system
Ay is pronounced as the sound you (or at least I) make when pinched. In its first version, it was geared toward eye surgery clinics. In later versions, Ay was split into systems; Ay 2.0 specialized in holistic management and patient scheduling for medical practices and small hospitals, and the Dr. Eye offshoot that specialized in optometry.

Data Center

Design and construction of a global data center
For Corporate Express I led the design and construction of a global data center. This small center hosted our business systems, web hosting, and VOIP. The data center serviced the US headquarters, branches in Canada and Europe, and off-sites sales force. Beyond legacy machines and systems it was based on Dell servers, Microsoft Windows servers, SQL server, CISCO switches and appliances, SONET ring for complete survivability, and EMC for storage and replication. The St. Louis based data center was paired with a disaster recovery site in Denver.

Image Compression

Machine Learning Image Compression
Using Machine Learning for Image Compression.
 

Credits

I have collaborated with many creative and talented people. Being surrounded by top talent and very, very smart people is always my first goal in every project. There is great satisfaction in learning from others. Unfortunately I may have missed a few great people in this section. Send me a note if that misfortune did occur. I developed this web site (this was a cool project in and of itself) using the best technology available and it is very easy to update.

I did not invent the FPGA-based deep packet inspection appliance. The inventor is Greg Sullivan to whom I will always be in debt for introducing me to his endeavors. Together we have invented numerous futuristic technologies and addressed federally sponsored calls for white-papers (none of which can appear on this site). I enjoy every discussion I have with this creative leader, whose ideas often come before their time. I became involved with the FPGA appliance development late in the cycle, but I have had the opportunity to lead the transition from this proprietary hardware based system to the novel Securio cloud.

Packman AI and VLSI 8-bit RISC CPU were not forged in my brain either. These two are graduate-level projects that were long, challenging, and product-like-complete, surely wordy of being mentioned here. I also immensely enjoyed working on them and this is my site, so here we go, they are here. The Packman AI as a vehicle for teaching and the whole infrastructure development is due primary to Dr. John DeNero and Dr. Dan Klein, both at Berkeley University. I just happened to be in the right class in the right time, and lucky enough the join the right team. Peng Li, Daniel Lazewatsky, and Steve Cole; it was a pleasure and intellectually stimulating working with you. For more details about where Packman AI is now (if it was not eaten) click here.

The VLSI 8-bit RISC CPU is credited to the great Dr. Viktor Gruev. I wish I had more opportunities to work near his brilliance. His lab is constantly inventing new science and I highly recommend following their work.

The SOA project idea is credited to Dave Brown. This project drastically changed the company we worked for and he envisioned it when all people knew about SOA was that it was a TV show.