|BIOGRAPHY & IMAGES||2|
|PATENTS & PUBLICATIONS||4|
BIOGRAPHY & IMAGES
Brian Helfrich has more than 25 years of experience in computer technology research and development. He's currently Chief Architect at Helcorp Interactive, a company in Seattle developing a machine learning system. Prior to Helcorp, he worked at Amazon as a Software Development Manager in the Amazon Android Appstore where he managed the development and operations of services in support of the Kindle Fire and Android mobile devices. Prior to the Appstore, he managed the development of Amazon's EC2 Spot Instances, a spot market for selling cloud computing resources. Before landing at Amazon, he served in multiple technology leadership roles in high-tech startup and R&D companies in a variety of industries including machine learning, gaming, mobile, cloud computing, search, social media, computational nanotechnology, speech recognition, and marketing. Brian started his career at Telcordia Technologies (formerly Bellcore Labs) where he was a Research Scientist working on automated speech recognition applications, visualization of complex systems, and new business development. Brian holds a BA degree in Computer Science from UC Berkeley. He has seven patents, including the patent for the pricing algorithm used in Amazon's EC2 Spot Market.
SUMMARY OF QUALIFICATIONS
- Creative systems architect and software development engineer drawing from various professional and personal experiences to devise scalable and efficient solutions for difficult and complex problems.
- Highly proficient in systems architecture and integration, object-oriented programming, distributed computing development, and scientific computing and visualization.
- Over 25 years of experience serving multiple roles in high-tech startup and R&D companies in various industries including machine learning, gaming, mobile, cloud computing, search, social media, computational nanotechnology, speech recognition, and marketing.
- BA, Computer Science, University of California at Berkeley (1999)
Chief Architect - HELCORP INTERACTIVE, Seattle, WA (1/14 - present)
- Ongoing research and development of machine learning systems involved with automated threat detection, human mimicry, thought, and planning strategies utilizing consciousness, self-awareness, sentience, and sapience models
- Architecture and development of a real-time, automated system for coordinating multiple player agents towards completion of real-world missions in an alternate reality game
- The system supports GPS locations, SMS, mobile client interactions, redacted redacted redacted inputs, and communicates with a mobile client
- Created a mobile client which communicates with the above coordination system, and provides audio, text, images, video, and dynamic map visualizations to assist the player in the completion of their mission
- Created an inventory management app for preppers featuring an onboarding workflow to generate prepping goals and track item expiry [LINK]
- Responsible for the architecture, development, and operations of
- Amazon Android Appstore (1/13 - 11/13)
- Amazon Web Service's EC2 Spot Instances (4/12 - 1/13)
- Architecture and development of Amazon Web Service's EC2 Reserved and Spot Instances products
- Created the statistical model and algorithm of the next-generation spot pricing system for EC2 compute resources
- redacted redacted redacted redacted redacted for redacted redacted redacted redacted> and redacted redacted redacted redacted>
- Designed and implemented a Sphinx-based search engine which indexes status
messages and share contents of user streams
- Updated with global rank velocities and user engagement to enable sorting search results by popularity
- Created a prototype document topic extraction system blending results from TF*IDF, NER, NP-chunking, and OpenCalais
- Created a distributed task scheduler designed for social network tasks such as fetching user streams and inbox messages at API rate limits
- Integrated and formalized the company's automated deployment and monitoring
system for computer instances and software
- A combination of EC2 control scripts, Chef, Nagios, and Ganglia packages
- Designed and implemented an agent-based distributed computing system used to
process and store over 100 million documents producing over 400 million objects
(text files, images, and PDFs) consuming 100 terabytes of storage.
- The system continually ingests raw data from various sources ranging from tapes sent from the US patent office to live press release data pushed to collection points.
- The system's agents are fully scriptable and operate in an environment providing text processing, graphics, OCR, named entity recognition, and natural language processing services.
- The system runs on Amazon Web Services with load-sensitive dynamic resource provisioning for efficient utilization of pay-per-hour compute instances.
- Created "Boliven Web Services" comprising
- A set of web services providing RESTful API access to Boliven's data for use internally, and externally on a pay-per-call basis.
- A set of widgets (patent search/PDF boxes, patent drawings scrollboxes, scrolling data feeds) and RSS feeds used to syndicate Boliven's data and functionality to external sites.
- Coordinated the design and execution of a system to pre-calculate the social network of people implied by their appearances on over 100 million documents.
- Created and managed a project to find failure modes for 25 published carbon-
dimer mechanosynthesis tooltips. Implemented a ~10,000 node, globally dis-
tributed computing network to run the project (nanohive-1.org/atHome) and used
GROMACS to generate tooltip conformations for quantum chemical minimization
- The goal of the project was to find failure modes for 25 published C2 dimer tooltips.
- The approach was to perform a molecular dynamics simulation of each molecule at a very high temperature so that it deformed and undulated significantly, then to perform quantum chemical geometry optimizations on each deformed variant produced in the simulation. This approach mechanized the creation of initial, unstable structure variants that could then minimize to alternative structures.
- Designed and implemented an HDF5-based simulation results library and int- egrated it within GROMACS as a new format for writing simulation results data such as trajectories and energies.
- Prototyped the integration of GROMACS with NanoEngineer-1 (a molecular chem- istry CAD tool) for atomistic and reduced model molecular dynamics and energy minimization of DNA.
- Lead the design and development of NanoVision-1, a cross-platform, Qt-based application that displays chemical structures and simulations especially those produced by NanoEngineer-1 and GROMACS.
- Designed and implemented a modular, cross-platform, simulation framework for molecular modeling. The architecture supports plugins for simulator control, data import/export/storage, multiscale time/space traversal, distributed computing, physical interactions, and simulation results.
- Created plugins for
- Simulator control via Java, TCP sockets, and Python.
- In-lab distributed computing (DC) via sockets and @Home-style DC via the Berkeley Open Infrastructure for Network Computing (BOINC).
- Integration of Reactive Empirical Bond Order (REBO) molecular dynamics, and ab initio methods such as Hartree-Fock, Kohn-Sham DFT, and Møller-Plesset perturbation theory.
- Simulation results via netCDF data files and POV-Ray scene descriptions.
- Created NanoHive@Home, a distributed computing system used for large-scale nanotech systems simulation and analysis that draws its computing power from otherwise idle computers sitting in people's homes.
- Research and development of OpRisk's operational risk severity model and fitting algorithm: supporting algorithm library research, integration and refinement of prototype code, graphical search space visualization, model testing and validation, and code performance and stability improvement.
- Designed and led the implementation of a fully distributed computing environ-
ment (DCE) for use in pre-processing the national telephone directory for use
in a speech-enabled directory assistance application.
- DCE features include sophisticated job specification, fail over and retry mechanisms, graphical control and monitoring interface with login-controlled access levels, SNMP messaging, and an extensible framework.
- Managed a team of five programmers for the implementation of the DCE.
- Managed the QA department and developed tools and techniques for testing the performance of a speech-enabled national directory assistance application.
- Co-Inventor of patent #6671670 - System and method for pre-processing infor- mation used by an automated attendant.
- Met with customers to explain ideas about how web, and related technology could help them achieve their marketing objectives, then later to discuss design and implementation details and planning.
- Collaborated internally with sales associates, media production personnel, and graphic artists.
- Managed a team of (3-5) programmers to deliver web promotions and sites ranging from referral programs to sophisticated on-line shops complete with secure product purchase, product search and reviews, affiliate programs, membership reward point system, gift registry, and configurable communities.
- Maintained a multi-platform network (Solaris, Linux, NT, Mac OS X) running Apache web servers, Oracle and MySQL databases.
- Spun-out Telelogue, a speech applications startup: prototype/demo development, product definition, and business plan.
- Re-tooled Telcordia's Intelligent Service Peripheral platform to support distributed speech-enabled applications.
- Prototyped a graphical monitoring tool for Telcordia's large-scale workflow management system.
- redacted redacted redacted redacted with redacted redacted redacted redacted redacted redacted redacted> for redacted redacted redacted>
- Created the user-model, user-interface, and user-description to system-call compiler for creating and deploying Internet-based workflow.
- Wrote an automated seminar/conference management system capable of JavaOne- sized events.
- Wrote a data-validation and de-duplication system for web-based lead gener- ation and management.
- Performed a heuristic usability evaluation of the complete system.
- Wrote a low-level system expression evaluator.
- Developed prototype of an expert system used to minimize burr formations for metal part fabrication. (VB, C++, ODBC)
- Consulted with Yahoo! website engineers to write custom web log processing scripts to complete monthly audits of their then million-visit-per-day site.
- Wrote code libraries for processing incoming web log files.
- Rewrite of existing on-site web log transfer utility to use ftp and add more robust error recovery and retry mechanisms.
- Wrote a graphical ftp and web server uptime utility.
- Devised test plans and test harness for client/server web log analysis systems.
PATENTS & PUBLICATIONS
David Ward Jr., Jin Seop Kim, Stephen Elliott, Rusenia Ingram, Rachid Ennaji, Brian Helfrich, Dhanvi Kapila, Peng Zhai, Notification of computing capacity status, US Patent 10,163,152. 25 December 2018. [PDF]
David Ward Jr., Brian Helfrich, David Nunnerley, Cloud resource marketplace for third-party capacity, US Patent 9,760,928. 12 September 2017. [PDF]
Rachid Ennaji, Jin Seop Kim, Brian Helfrich, David Ward Jr., Stephen Elliott, Peng Zhai, Dhanvi Kapila, User-defined pools, US Patent 9,645,840. 9 May 2017. [PDF]
Rachid Ennaji, Jin Seop Kim, Brian Helfrich, David Ward Jr., Stephen Elliott, Peng Zhai, Dhanvi Kapila, Burst capacity for user-defined pools, US Patent 9,634,958. 25 April 2017. [PDF]
David Ward Jr., Geoffrey Pare, Brian Helfrich, David Nunnerley, Job resource planner for cloud computing environments, US Patent 8,676,622. 18 March 2014. [PDF]
Brian Helfrich, David Nunnerley, System and method for managing requests for pooled resources during non-contention, US Patent 8,676,621. 18 March 2014. [PDF]
Damian G. Allis, Brian Helfrich, Robert A. Freitas Jr., Ralph C. Merkle, Analysis of Diamondoid Mechanosynthesis Tooltip Pathologies Generated via a Distributed Computing Approach, J. Comput. Theor. Nanosci. 8(July 2011)1139-1161. [PDF]
Helfrich, B. NanoHive@Home - NanoSystems Simulation. Invited chapter, Distributed & Grid Computing - Science Made Transparent for Everyone. Principles, Applications, and Supporting Communities, 2008, edited by M. Weber, Rechenkraft.net, Marburg. [PDF]
Esther Levin, Susan Boyce, Brian Helfrich, Amir Mane, Alison Schondorf, Ilija Zeljkovic, System and method for pre-processing information used by an automated attendant, US Patent 6,671,670. 30 December 2003. [PDF]
Helfrich, B. and J. A. Landay, QUIP: Quantitative User Interface Profiling. 1999. [PDF]
Member, Nanofactory Collaboration [LINK]
Former Chef, Caribbean Zone, San Francisco [LINK]
RELEASE NO: 68-54k