Work with engineering, R & D and executive team to add capabilities to Vizit’s visual intelligence platform

- Add new features and do maintenance on back end machine learning / data processing code written in Scala- Developed new Python application for executing machine learning models (execution engine)- Technical lead of Execution Engine team

In my new position I will be working in the Amazon Privacy Group to build software and services to manage customers’ private data, ensuring compliance with privacy regulations in all the countries in which Amazon operates

- Implement AWS lambdas to break up monolithic application for greater scalability- Write Python code to ETL logging data into Elasticsearch to support Kibana queries/failure analyses- Modify legacy C# code to add new features and correct defects- Performed analyses in support of business needs on 3rd party software products, AWS costing

Enhance Bazaarvoice's lambda architecture based data management system:- Add new features to Bazaarvoice's streaming rules engine (Polloi), which is essentially a materialized view generator that generates "denormalized" JSON documents from a subscription (CDC stream) from Bazaarvoice's EmoDB database- Do prototype design and development of a new version of the Polloi software (“Polloi X”) that could output denormalized documents to Kinesis instead of Elasticsearch, present to the R&D division- Investigate changing EmoDB's CDC stream generator (Databus) to use Kafka to distribute the CDCs instead of Cassandra, build prototype based on Kafka Streams- Conduct parallel investigation and prototype development using Apache Pulsar- Document and present these two prototype designs and AWS hosted stress test results to engineering teams in Bazaarvoice R&D- Develop training presentations on Polloi software internals and Bazaarvoice data pipeline, train other teams to maintain parquet file generation software for the data pipeline- Do maintenance programming on Polloi and EmoDB and related services

Develop back-end services code in Scala using functional programming techniques. These services will both query existing databases and transform unstructured data (e.g. search engine results) into structured data for populating databases to support complex queries used to gather information about corporations for sales lead purposes. • Performed maintenance programming in Scala on a master data management (MDM) web service used to synchronize Cisco Inc.’s customer relations management (CRM) database with the Avention database• Developed new streamlined MDM software baseline using declarative validation and transformation for JSON on REST APIs that would support multiple customers without changing code• Developed new transforms and associated SQL and Javascript scripts for Pentaho ETL environment to automate the processing of customer bulk data

Performing maintenance programming in Java on ASPEN, a radar signal analysis tool suite that is used by technical electronic intelligence (TECH ELINT) analysts to discover radar system capabilities and functions.Designed GUI front end as well as server-side back end for a radar scan simulator as part of a new ASPEN radar scan analysis and modeling tool.

Designed and developed software modifications in C++ (QNX RTOS, RTI DDS) for the AN/TPQ-50 radar to enable it to perform simultaneous target tracking and air search missions. Designed and developed a Java persistence mechanism (High-performance Persistence Framework or HPF, patent application in process) for mining large volumes of electromagnetic data for signals of interest. Resulted in much higher performance than previous approach based on commercial relational databases.Designed and developed Java chemical classification software for performing analog identification matching (AIM) to classify a new chemical against an existing database of chemicals to determine potential carcinogenic or toxicity hazards. Reduced analysis time from days to fraction of a second, presented this work to the analytic chemistry community at QSAR 2008. Led to SRC being awarded the ChemACE contract from the Environmental Protection Agency (EPA).Authored requirements for, designed and developed electromagnetic statistical evaluation tools in Java (Windows) for detection of threat signals in battle theaters. Reduced analysis time from days to hours.Designed and developed cyber data mining software in Java (Mac OS X) based on HPF for the Air Force Computer Emergency Response Team (AFCERT) at Lackland AFB. Time to detect intrusions, exfiltrations, etc. reduced from 24+ hours to seconds. Successfully deployed turnkey system to AFCERTDesigned and developed Domain Specific Language (DSL), the Radar Signal Description Language (RSDL), for modeling complex radar signals. Designed and developed the Concurrent Classification and Deinterleaving Machine (CCDM) software in Java, a novel state-machine based approach for concurrent emitter deinterleaving and classification as well as simulation (patent awarded)Received Several Achievement Awards and an Inventor’s Award

Designed and developed new real-time software in C++ (Embedded Linux) for next-generation prototype of the Airport Surface Detection Equipment Model X, or ASDE-X, multilateration systemDesigned and developed R*-tree based spatial data management software in C++ (Solaris) for use in air traffic control Traffic Information Systems Broadcast (TIS-B) demonstration softwareDeveloped real-time message processing software in C++ (VxWorks) for Alaska Capstone Automatic Dependent Surveillance Broadcast (ADS-B) air traffic control systemPerformed time series data filtering research for medical applications and developed SCADA software for water quality monitoring system

AN/SLY-2 Program (9/1999 - 10/2001)- Software team leader for the Emitter Classification (EC) software for the AN/SLY-2 surface ship electronic warfare (EW) system. - Developed new emitter classification software in Ada95 and C++ (VxWorks PPC) based on pulse-to-pulse modulation, became the systems engineer for the EC software and wrote the software specification for the classification subsystem- Worked with EW analysts to develop optimal spatial indexing strategy based on R* trees for emitter signal models from the system threat library- Modified FIRM object database engine for use in this system to manage emitter threat library and system track file within real-time constraints

FIRM Program- Software team leader for the Functionally Integrated Resource Manager (FIRM), a next-generation real-time object database program run by the Air Force Research Laboratory (AFRL) at Wright-Patterson AFB, intended for use in F-22 avionics- Developed concurrency control software, internal data structures, and top-level API specification and code in Ada95- Represented my employer as a voting member of the Object Data Management Group (ODMG) in development of an Ada95 binding to the ODMG object model- Co-edited object model (Chapter 2) of ODMG standard- Represented Lockheed Martin to military standardization groups (Real Time Architecture Group (RTAG), Database Interface Standards Working Group (DISWG))

AN/BSY-2 Seawolf Combat System Program- Developed high-speed indexing (AVL tree) software in Ada for the real time database manager (RTDM) database engine for the AN/BSY-2 (Seawolf) combat system- Designed and developed persistent storage software for RTDM in Ada, including threaded B*-trees for optimal disk-based data indexing- Developed caching algorithm for persistent data access, added code for it to the RTDM software- Gave numerous presentations on the RTDM software to the customer, developed test plan to ensure deterministic performance under all loading conditions, conducted formal record run real-time performance tests with Naval Undersea Warfare Center (NUWC)- Assisted in development of commercial version of the RTDM software

AN/FPS-117 Seek Igloo program and AN/FPS-118 OTH-B program- Developed multi-scan correlator software for the AN/FPS-117 radar (Seek Igloo program)- Developed signal processing software for AN/FPS-118 (Over The Horizon Backscatter, or OTH-B) radar that would allow the system to track ships as well as aircraft. This required development of special data structures and algorithms to process near-zero Doppler data in high noise (Bragg line) conditions- Developed improved tracking algorithms for AN/FPS-117 radar- Graduated from Edison Engineering Program and Advanced Course in Computers during this time while also earning a MS in Computer Engineering from Syracuse University