Continuing my work on signal processing, automatic speech recognition, and machine learning from Nuance now as part of Microsoft Cloud and AI.

Software developer and researcher working on real-time multi-channel audio processing for machine learning-based medical report generation system.* Worked with researchers to take Python prototype beamforming algorithm to real-time C++ production system in only 3 months.* Developed GRPC/Protobuf system to transmit on-premises audio to cloud-based machine learning system.* Researched signal correlation and adaptive filtering algorithms for use in improving signal to noise ratio in captured audio.Software developer working on speech middleware used by nearly all of Nuance's many speech recognition products, from local recognition on hand-held devices, to Dragon NaturallySpeaking on PCs and Macs, to Linux-based recognition in the cloud. * One of two designers who completely rearchitected speech middleware software, integrating a full scripting language and providing a fast bridge between leading-edge research and production systems.* C++ and Python development for recognition accuracy tests on Nuance's supercomputing "grid."* Led the architecture and implementation of both server and local recognition engine speech adaptation processes for Dragon Assistant.* Developed speech recognition logic for Dragon 12 and 13. Primary technical contact for my team during Dragon 13 program.

* Joined the Media Engine/"Player" group and led the implementation of a new, asynchronous method of operating the video and special effects pipeline, as well as supporting the Avid Nitris hardware platform in the Media Composer product line. This culminated in the release of Symphony Nitris in 2005.* Directed and helped to implement many architectural improvements in the Media Composer/Symphony media engine, including such features as the integration of the DX hardware product line, support for GPU effect processing, ongoing improvements for multi-core scalability, and the addition of an entirely new pipeline for ancillary data.

* Analyzed and optimized the special effects pipeline for Media Composer, finding bottlenecks and devising a patented architecture that provided significant performance improvements for software effects. * Worked with Avid's hardware design team to implement a "safe color limiter" capability that provided smart clipping to the NTSC composite color space, in real time.* Worked on the design and implementation of the Symphony Color Corrector, including several novel features for which we were awarded patents. I was co-inventor for three patents on this project.* Led the implementation of a new hardware effect subsystem for Avid's Nitris hardware platform, culminating in the release of DS|Nitris.

Wrote a wide variety of software for the special effects group for the Media Composer/Symphony product line. This included driver-level and user-level code for a 2D special effects board developed in-house, as well as user-level 3D special effects code for a 3rd-party board.

Sole software engineer for the internal R & D group in the Avionics Division. Worked with a small team of hardware engineers to develop real-time graphics boards to display moving maps and process live video. Wrote all software to demonstrate these boards, from low-level device drivers and interrupt handlers, through demonstration applications and simple command-line user interface logic. Demonstrated texture-mapped graphics and video textures before these terms became commonplace in the computer graphics community.Held Secret clearance.