Leading machine learning R&D, focused on highly accurate recognition of Linux processes exhibiting malware behaviors, thereby enabling lightening-fast identification and remediation of zero-day threats. Providing customer-facing support and solutions to ensure our users are happy and protected.

Continuation of role as Language Services Lead.Invented and implemented a novel method of automatically generating "substitution" rules that improve ASR accuracy in general and substantialy improve F1 scores for words with high analytic value, including out-of-vocabulary (OOV) words. Completed over 50 accuracy improvement projects for customers using this method.Invented and implemented a method of using Presidio Named-Entity-Recognition (NER) text tagging to redact Personal-Identifying-Information (PII) and sensitive numbers from speech audio with word-level precision, to protect customer privacy and comply with Payment Card Industry (PCI) regulations. Implemented as Docker image serving a REST API to clients.Completed several full-stack ASR modeling projects, including training and customization of acoustic, language, and substitution modeling layers. Full-stack modeling improves overall recognition accuracy for whole industries and world regions.Developed prototype software that uses forced alignment technique to add word-level time-stamps to OpenAI Whisper transcripts. Prototype also removes non-speech audio regions to reduce the impact of Whisper's hallucination issues and other text processing methods to make the text more useful for down-stream processes. Implemented as Docker image running several Whisper instances in parallel to maximize individual GPU utilization, plus many instances of these images running in parallel assigned to different GPUS, thereby maximizing overall server capacity utilization.Developed test frameworks for evaluating new ASR features and releases.Supported sales by providing benchmarking and POC services on short notice as needed. Continued mentoring junior engineering staff.Created numerous documents using Google and Atlassian collaboration tools as well as MS Office and and Apple Mac tools. Can easily switch between Apple, Microsoft, and Linux development platforms at will.

Focused on ensuring that our customers get the best possible accuracy from Voci's ASR products. Responsible for applying a spectrum of NLP and machine learning techniques to provide the greatest benefits possible within each customer's cost and schedule constraints. The role includes significant software development to automate machine learning training flows and continuous improvement of ML models and substitution rule sets.

Focused on delivering exceptional customer experiences with Voci's portfolio of industry-leading speech transcription, identification, and analysis solutions. Responsibilities include software development, CentOS Linux administration, systems integration, troubleshooting, mentoring, training, and technical documentation.

Focused on delivering exceptional customer experiences with Voci's unique speech analytics offerings. Responsibilities include quality assurance for software products written in multiple languages, development and enhancement of integration software, engaging with prospects and customers to ensure the smooth operation of Voci's technical products, and creating and enhancing written documentation (both internal and customer-facing).

Responsible for maintaining and applying standards for safety and quality across all RDT&E Lab operations, including individual projects. Provide team leadership and directly manage technical, schedule and budgetary aspects of especially challenging projects/programs undertaken by the Lab. Support business development efforts, both government and commercial. Continued direct technical development in areas including embedded system design and algorithm development/implementation as needed.

Collaborated with a strong research team to devise novel solutions to meet the needs of our customers, including Federal Government agencies, commercial entities, and non-profit organizations. Contributed to business development and market analysis activities. R&D contributions were primarily in the areas of embedded systems, signal analysis, and automated testing.

Lead the Coastal SC Section of the IEEE. Hosted technical meetings of interest to our members, including bringing distinguished lecturers in from around the country to share their knowledge and insights. Keep up with the Section at http://www.ieee.org/go/coastalsc.

Assisted with coordination of technical meetings for the Coastal SC Section of the IEEE. Provide meeting facilitation support and other services as needed to effectively market and accomplish Section activities.

Developed automated test software for Software Defined Radio systems. Performed development and testing work at SPAWAR, Charleston SC facility.

Developed web-based database applications to support multiple programs. Also served as Facility Security Officer and supported business development efforts.

Continued NeoCell development after Cadence bought Neolinear in 2003. Developed an automated test environment using grid-computing tools to test a non-deterministic software product. Used statistical techniques on large number of test results to identify NeoCell behaviors that needed improvement or that had subtly regressed as a result of an algorithmic change.

Developed Electronics CAD (ECAD) framework at the heart of a novel analog integrated circuit physical synthesis automation tool called NeoCell. This tool was developed in association with Carnegie Mellon University.

Designed and implemented a data collection, monitoring and analysis hardware/software system for an experimental Plasma-Arc Chrome Smelting Furnace. Operated furnace in collaboration with MIT.Performed test and evaluation of the proposed ISO 10303-210 STEP standard for storage and translation of electrical product data (AP210).Developed and published a novel CAD data translation architecture called EXPRESS-Driven Data Translation (EDDT). Demonstrated EDDT by implementing a Mentor Board Station to AP210 (M2A) translator.Integrated M2A translator into the RASSP Manufacturing Interface which was subsequently used to substantially reduce cost and improve yield of Printed Circuit Assemblies (PCAs) produced by Lockheed Martin's manufacturing facility in Ocala, FL.