Took data from a wide range of sources and used React/JavaScript and KDB+ to analyze for system health, performance, security, failure prevention, etc. from Unix Syslog, Datadog, Splunk, Security Event Manager, etc. Developed a comprehensive data analysis methodology for system performance assessment, leading to a 20% improvement in efficiency through bottleneck identification and process streamlining. Analyzing customer behavior for business insights: Identified four key points at which an e-commerce website was sometimes slow and reduced abandonment 35%. Also, A/B testing analysis to find best product descriptions for certain client demographics. Security Enhancement: Analyzed system logs to identify and mitigate security vulnerabilities, speeding up the process of prioritizing and addressing a wide range of security issues.Used sklearn’s SVM package to determine which situations required focused human attention. Developing a system to support the work of security engineers designed to improve their efficiency by reducing manual workload. Use of KDB+ for reporting and Angular for UI.

Responsibilities: • For CJF, an organization with a long and unique history of successfully suing state sponsors of terrorism. Used KDB+ to analyze a range of data. • Used Python scikit-learn to help lawyers find documents of interest from large repositories during pre-trial discovery. • Wrote patent application for LLM-based system including demonstration code in Python. • Used Python NLTK to perform sentiment analysis on thousands of emails per day. • Used Python’s scikit-image to extracted maps and related images from large document libraries. • With Jenkins, automated various DevOps tasks including automating a pipeline that takes new datasets and runs them though a series of classifiers and distributes reports.

Responsibilities: Developed a set of proposals in the area of conflict reduction and thereafter built a publication infrastructure that allows for highly flexible publication including hosting a video transcoded into a wide range of formats, blogging, and secure private web analytics. Developed assets including games designed to help reduce the chances of young people falling victim to hate and other thinking errors.

Responsibilities: • Full-stack sole developer of the Learning Album Dashboard (LAD) system that monitors internal training/education for the largest U.S. wireless cellular provider, nationwide, across many functions including sales and technology operations. LAD, which displays many graphs (D3) and tables, is a critical decision support system that allows courseware developers, trainers, and their managers to see which courses are providing value and which need to be upgraded, what course modules are confusing, and why. • LAD was built as a highly flexible single-page app using HTML5/CSS, Node.js, Backbone / Marionette.js, Grunt, KDB+ / q from Kx Systems (a high-performance, high-volume database designed for analysis of vast amounts of data, most often used on Wall St.) and Linux. It employs complex server-side aggregation approaches to achieve high-performance data analysis and implements complex measurements on a rapidly evolving set of courseware assets. Newer version is in React.

• Full-stack sole developer of the Learning Album Dashboard (LAD) system that monitors internal training/education for the largest U.S. wireless cellular provider, nationwide, across many functions including sales and technology operations. LAD, which displays many graphs (D3) and tables, is a critical decision support system that allows courseware developers, trainers, and their managers to see which courses are providing value and which need to be upgraded, what course modules are confusing, and why. • LAD was built as a highly flexible single-page app using HTML5/CSS, Node.js, Backbone / Marionette.js, Grunt, KDB+ / q from Kx Systems (a high-performance, high-volume database designed for analysis of vast amounts of data, most often used on Wall St.) and Linux. It employs complex server-side aggregation approaches to achieve high-performance data analysis and implements complex measurements on a rapidly evolving set of courseware assets. Newer version is in React.

For the Baikal/Levy partnership, developed a range of computer applications including 3D hyper-realistic models of custom leather goods including handbags. Used Node/HTML5. Used Python scripts to scrape data from a range of websites and for other tasks.

Full-stack developer in charge of all technology of the one-stop-shop / largest database of private company deals available anywhere, targeted at the accredited investor and venture fund. Defined the overall architecture (using AngularJS, RabbitMQ, MongoDB, Gulp.js and Elasticsearch with MongoDB used as fast session store.) and the initial user experience. Addressed such key technical challenges as achieving extremely high scalability and super-fast sorting and searching. Oversaw the development of a high-tech breakthrough recommendation engine for matching deals to investor interests. Took the company from no technology at all to launch and beyond. Worked on AI-centric upgrade, hiring, and managing all key technical staff. Lead a team of 10 developers as well as DBAs and admin staff. Bootstrap for responsiveness.

Built single-page apps using HTML/CSS with jQuery, Node.js, Backbone / Marionette.js, and Grunt for delivery of election data. System was in use in the 2015/16 election cycle by reporters, editors, analysts, producers, etc. during primary and general elections. Built server and client-side elements. System used successfully during the primaries and general election in 2016 and will continue to be used both in practice mode and election mode in 2018 and thereafter.

Oversee the largest election threat model validation effort ever undertaken, collaborating with researchers at top US and international Universities, security experts and other relevant, specialized organizations. Working under Stanford University / National Science Foundation contract. Goals include improving allocation of election resources by evaluating and rating threats to election integrity.

Methodology Creation for Election Attack Threat Analysis and Security Enhancements. Developed proprietary threat-analysis application, “AttackDog,” employed at the National Institute for Standards and Technology (NIST, US Commerce Department) by the team that writes voting-system standards. Developed a repeatable, software-supported method for allocation of security-related resources incorporating probabilistic mathematical methods. Can be used to prioritize attacks and estimate the effectiveness of security measures. Collaborated with Stanford Professor David Dill. Funded by the NSF.

Directed a team of 17 technology experts and outside election officials to evaluate the plans to move one of the nation's largest and most complex election operations to vote-by-mail in time for the 2008 general election. Acquired and deployed millions of dollars worth of mail sorting and vote tabulation hardware and software.

Lead project within Brennan Center for Justice of New York University involving 40 multi-disciplinary experts. Developed process for assessing the relative power of security measures. Cataloged over 100 attacks against the three main sorts of polling-place voting-systems in current use. Assembled the mix of talent to create a team of leading computer-security experts and coordinated with election officials to develop assessment of risks of fraud and error. Created and nurtured strong alliance with NIST that lead to coordinated publication. Report cited as “historic” by famed MIT cryptographer and RSA inventor Dr. Ron Rivest, achieving wide-spread media coverage.

Created a hosted solution to support the Structured Settlement Acquisition Industry. Lead a team to construct a work flow/data warehouse combination to automate the acquisition and resale of structured settlements and flexibly report on deals as they move though departments of the client firm. PAX reports provided on-going, customized views about deals for critical business judgments and prompt decision making. Responsibilities included managing all aspects of development including team recruiting, requirements analysis, scheduling and fault prioritization. Performed hands-on requirements analysis, architecture and database design.

Developed a productivity software suite facilitating high-productivity group problem solving. Suite supports related approaches to group problem solving, including Work-Out (used at GE) and GoFast (used at GM). Managed four developers and five domain experts and filed patents.

Generated recommendations on voting technology for the November 2004 elections which were endorsed by many of the leading figures in computer security. Created consensus where there had been a fractious debate. Managed team of 12 to develop recommendations for the civil rights community for the Brennan Center for Justice of New York University in alliance with the Leadership Conference on Civil Rights (LCCR).

Managed the development of a powerful user interface framework that empowers programmers to manage the complexity of web software development, maintain a clear separation between presentation functionality and underlying business logic and to support complex transactional semantics. The software utilizes the Model View Controller pattern, JSP Custom Tags and Servlet technologies. This tool achieves a six-fold programmer productivity gain over conventional technology during both development and maintenance activities. Producing a dramatic reduction in code bulk, WebTier runs under Oracle, BEA, WebSphere, Tomcat and Orion, among others. (Patent issued May 2006.)

Designed and managed the build of major portions of a powerful, interactive system controlling compliance for financial instruments, including futures, caps floors and more exotic structures. The software product was based on J2EE technologies, including Enterprise Java Beans (EJB) Servlets, JSP, JNDI and JMS. Development and test environments included IBM WebSphere and WebLogic Application servers, Sybase, Microsoft SQL Server, Oracle Databases and Windows and Solaris operating systems. Designed new work flow engine, new representation of innovative financial products and a pluggable architecture for extendibility. Oversaw deployment at Deutsche Bank, Citibank, Credit Swiss/First Boston, Chase and other financial institutions.

Designed a content management system in use at General Motors Corporation for data analysis. Managed all aspects of this breakthrough technology facilitating real-time collaboration among statisticians, data mining experts, other data analysts, marketing and sales executives.

Designed systems to automatically generate and manage retail prices by executing customizable, declarative business rules. Used Java, Swing, CORBA, XML, SQL, Oracle 8, C++, MS Windows and Solaris to develop a highly flexible and distributed application. System was considered a generation jump over other available industry solutions and was deployed at some of the largest retailers in the world, including SuperValu. Managed 60 programmers and seven analysts and requirements specialists.

Developed a sophisticated and innovative application transforming the sales and marketing function. Led a global team of 14 developers, three requirements specialists and external technologist consultants on two continents. This graphics-intensive, three-tier client/server system was deployed on laptop computers in the field, accessing a multiple gigabyte data warehouse running on a multi-million dollar symmetric multiprocessor. Management scope included: (1) leading the development of the front-end: analysis and design, (2) developing the communications framework, (3) coding some of the most complex elements of the system, (4) mentoring employees in the use of Smalltalk and version control, and (5) planning the testing packaging and delivery.

Managed four LTM developers in the restructuring of a large, client/server system. Employed Smalltalk and Sybase. Management scope included: (1) developed new, reusable framework architecture, (2) implemented new architecture prototype, (3) moved over 40 windows into the new architecture, (4) trained/mentored Prudential developers in the use of OO methods, Smalltalk and Smalltalk tools and effective use of version control tools. Shrank code size substantially increasing maintainability and efficiency.

Developed a system that tracked millions of dollars in promotional spend for a major manufacturer of 19,000 products using object and relational technologies. Trained staff in object-oriented technology.

Wrote course materials and trained developers in object technology at Chase Manhattan Bank, The Institute for Software Advancement, Mass Mutual, Shawmut Bank, Uniroyal Chemical Co, and NYNEX Mobile. Trained non-technological employees of The Registry (a consulting firm providing IT services for Fortune 1000 companies) in the concepts of object technology, enabling them to assist development managers.

Rapidly prototyped a variety of microcomputer applications using InterLisp-D, an early window-oriented workstation programming system. Developed presentation-graphics package for commercial deployment.

Led initial implementation of the configuration system (plus associated knowledge acquisition tool) for IBM's entire hardware product line. The system, deployed on two types of workstations, is used in knowledge acquisition and testing at three sites and two countries. Consulted in development of underlying, highly innovative, declarative, object-oriented, constraint-based language.

Built software re-engineering applications and created constraint-based development tool for the design of system integration requirements for end-client XEROX Corporation. Also conducted user-interface research.

Led a project in the management of operations for this leading supplier of software tools for artificial intelligence and knowledge-based systems applications development. Consulted for major corporations including Navistar, Texaco, Union Bank of Switzerland, 3M and Unisys. Trained teams from client firms, both domain experts and programmers, in the engineering of knowledge-based systems while building prototypes of potentially profitable applications. Wrote core modules of the first version of IntelliScope, IntelliCorp's radically new database front-end product. Wrote large parts of Strategene, a knowledge-based assistant for genetic engineering, marketed by AMOCO, including a user-interface graphics module.

Created systems architecture for JPMorgan's large, distributed, object-oriented trading application, Kapital, which was deployed on the derivatives desk at JPMorgan Chase in NYC, Tokyo and London.