Engineer in the Motion Software Team to work on algorithm verification and validation for next generation motion features. Working on sensor fusion, motion gestures, fitness tracking, etc.

Abbott Laboratories Diagnostics Division Hematology SiteSoftware developer for new product development under the R&D Algorithm team with a focus on computer vision and general algorithm support work. Creating inhouse tools to assist supervised machine learning tasks, automating data analysis and performance evaluation. Developed ML training and test review pipeline to shorten algorithm development cycles and integrate new training/testing data. Developed a streamlined Annotation Tool (.NET WPF application) to build up ground truth data with consensus across multiple manual reviewer domain experts.Designed and developed a full end to end domain specific data analysis dashboard. Full stack development of front and back-end web app using Python, Flask, Bokeh, Matplotlib, and Pandas. Highly configurable and flexible analysis interactivity spanning high level summary performance stats down into deep dive raw data low level inspection of outliers and classification anomalies for ML algorithm engineers to inspect results during prototyping and accelerate their development. Very powerful and well received data analysis and data visualization tooling supporting multiple views into 8-dimensional flow cytometry list-mode/timeseries data and the corresponding classification clusters.

• Summary: C++ developer for Image Lab™ image acquisition and analysis software used by quality control staff in the pharmaceutical industry as well as life science researchers from academia, government, and industry. Hands on experience with implementation of new features (Image Lab Security Edition), UI dialogs, build jobs, unit testing, documentation, bug fixing, and defect tracking.• Feature Development: directly responsible for bringing onboard user authentication functionality and security features for ensuring a secure application session for audit logging of regulated application commands. Such functionality is ultimately used to satisfy the FDA’s 21-CFR Part-11 regulation which stipulates various rules for maintaining, supporting, and controlling electronic data authenticity and integrity. Such functionality is critical to our pharmaceutical industry customers in their need to establish these practices in their SOPs. This development involved OO infrastructure design, Windows API and Active Directory programming, as well as intuitive UI creation within the Qt framework. Also developed a new scan coordinate UI for interfacing with our new flatbed gel and blot scanner the GS-900.• Development Environment: platform independent software production in an agile setting orchestrated around two week sprints holding much collaboration with very active SQA, product management, system integration, and marketing teams. Parallel team efforts managed through Perforce verified through continuous integration alongside unit, integration, and automated regression testing.

• Summary: C++ software developer for the team that designs and maintains the imaging engine of Siemens’ flagship 4D ultrasound system the Acuson SC2000™, a product hailed by cardiologists the world over• Feature Development: executed a large scale infrastructural refactoring to vectorize the codebase to support biplane imaging for the SC2000 instrument control software. This feature permits avoidance of exploratory 3D imaging and its lengthier scan times / lower resolution and instead permits physicians the ability to track two or more independent imaging planes to rapidly locate candidate anatomy before executing a 3D scan.• Tool Development: designed and implemented a software architecture visualization tool based on DGML and custom C++ software built to make tractable the comprehension of our massive data flow across our large codebase by way of a dependency graph representation spanning over 6000 nodes and 16000 edges.• Quality Control: conducted statistical analysis of system performance data to monitor the effects of new features; regular participant in code reviews and automated test report analysis for daily incremental builds.• Automation: applied diagnostic MatLab script for analyzing boot cycling screen shots; shell scripting and regular expression know how applied to: repetitive development tasks, locating bug prone regions of code, invoking mass automated refactoring via a sed script.• Development Environment: Windows platform software production in an agile setting orchestrated around FDA regulations for medical device software compliance (which translated to strict check-in guidelines and a paper trail of testing and development). Parallel team efforts managed through IBM’s Rational ClearCase, verified through incremental builds and subjected to unit/integration/regression testing.

• Designed an image based feedback trajectory correction algorithm for the gimbal mirrors in a MEMS confocal microscope device exhibiting visual artifacts. This method was based upon test images of a fine grating to sample trajectory points extracted from DAQ acquired time points corresponding to grating edge positions, thereby permitting trajectory reconstruction via spline interpolation. The resulting deviation from the ideal image was computed and fed back to the trajectory’s driving waveform for feed-forward correction.• Created an animated simulation of the scanning mirror operation in MatLab as well as conducting a parametric error analysis of involved components: grating dimensions, mirror oscillation frequency, interpolation method, and digitization sampling frequency.• Presented results at HHMI 2008 EXROP Annual Meeting