A passionate and innovative problem solver with a solid foundation in software and mechanical engineering, complemented by extensive experience in data management, system integration, user interface design, and machine learning. My academic journey culminated in a B.S. in Mechanical Engineering from the University of California, Santa Barbara, where I led award-winning projects. Professionally, I have a proven track record as a software engineer and architect at FLIR Systems, Inc., where I developed cutting-edge software systems to enhance production efficiency and accuracy in thermal camera manufacturing. My work spans from developing deep learning models for defect detection to creating automated systems for neuroscience research, significantly boosting productivity and research capabilities. I am committed to leveraging my technical expertise to tackle impactful challenges and create meaningful solutions.

I take pride in writing beautiful code and elegant software architectures. How a problem is solved often matters more than the problem itself.

As part of the Manufacturing Test and Automation team at FLIR, I spearheaded the development of an innovative machine learning algorithm designed to identify "blob" type defects within the manufacturing stages of thermal focal plane arrays (FPAs). Traditional detection techniques were proficient at identifying common defects like scratches, noisy arrays, or shorted/open pixels, but the challenge of detecting blob defects remained largely unsolved, leading to a reliance on manual inspection methods that were both time-consuming and inconsistent.

To tackle this issue, I initiated a project that leveraged a deep learning framework to autonomously recognize blob-type anomalies in FPAs. This solution was developed by training a model on the existing dataset of human-annotated images. The implementation of this model marked a significant technological advancement, achieving an impressive detection accuracy of approximately 95%. This breakthrough not only enhanced the efficiency of the manufacturing process by early identification and removal of defective FPAs but also generated considerable cost savings by preventing these units from advancing to more costly stages of production.

During my time at FLIR, I contributed significantly to the development of critical sensor technology for the Terminal High Altitude Air Defense (THAAD) system, a cornerstone of the United States' defense against Intercontinental Ballistic Missiles (ICBMs). THAAD is a system designed to intercept incoming ballistic missiles, including nuclear threats, during their terminal phase of their approach to impact. FLIR's cryogenically cooled mid-wavelength IR sensors are pivotal for the THAAD system's ability to detect and track incoming threats, ensuring a robust defense mechanism.

I acted as the primary software developer for the final Quality Assurance (QA) testing of THAAD sensors. This involved intricate software engineering to interface with complex electromechanical systems for data collection and analysis.
Our software served as FLIR's final validation step for the sensors, necessitating adherence to stringent standards for repeatability and measurement accuracy. I successfully navigated extensive third-party validation processes to guarantee sensor reliability in critical scenarios.

While my career at FLIR involved contributing to defense technology, I am deeply committed to ethical principles and the pursuit of work that serves constructive purposes. I recognize the importance of defense systems in maintaining peace and security, and I am proud of my contributions to THAAD. However, moving forward, my focus is to leverage my expertise in technology and leadership towards projects that directly contribute to constructive, rather than destructive, efforts.

In the Dr. Goard Research Lab, training mice to play video games was a critical yet slow and costly process, heavily reliant on human supervision. Identifying the need for innovation, I led a team of undergraduate engineers to develop a groundbreaking automated training device, tackling this challenge head-on. Our multidisciplinary approach resulted in a system that not only automated the training process but did so in a way that was engaging for mice, safe, reliable, and cost-effective. Key to our success was creating a user-friendly interface that allowed lab technicians without engineering backgrounds to easily maintain the system.

This project significantly advanced the lab's research capabilities by slashing training times by up to 80% and increasing our training capacity sixfold, thereby resolving the critical bottleneck of producing sufficiently trained mice for complex neuroscience experiments.

The impact of our work extends beyond just numbers; it has opened new avenues for research, leading to developments in the mapping of the visual cortex, and insights into how memories are formed.

We engineered a generalized game engine for similar projects, enabling the development of similar research games with high efficiency. For an in-depth exploration of our engine's architecture and capabilities, please refer to our comprehensive documentation: Game Engine Docs
To ensure seamless hardware integration and minimize response times, we also developed specialized Arduino firmware. This firmware transforms the Arduino Uno into an efficient, low-latency I/O device, crucial for our experiments' real-time demands. Discover the technical details and applications of our firmware here:Arduino Server Docs
Our project's source code, encompassing both the games designed for mouse training and the intricate hardware interfacing, is available for review. By sharing our code, we aim to contribute to the broader scientific and engineering communities, inviting collaboration and further innovation. Access our project's repositories at: Games and Hardware Interfacing, Game Engine

In a pioneering study with the University of California Santa Barbara’s Department of Anthropology, I embraced the role of a solo developer to create a suite of interactive video games. These games are designed to explore cognitive differences across genders, focusing on tasks related to hunting.

Discover the games
Immerse yourself in this research by downloading the games here. (Windows Installer x86)

Explore the Code
Delve into the development process and explore the source code on Github.

High Precision Pan-Tilt System for FLIR Cameras
As part of my Capstone project at the University of California, Santa Barbara, I collaborated with a talented team of engineers to explore a new potential drivetrain architecture for FLIR's pan-tilt lineup: Harmonic Drive®. This drivetrain technology eliminated the backlash that limited previous worm-drive systems, and aimed to bring unprecedented accuracy and repeatability into pan-tilt systems capable of seamless integration with FLIR’s thermal imaging cameras. This endeavor aimed not only to enhance the operational efficiency of FLIR’s imaging solutions but also to push the boundaries of high-zoom automated security monitoring systems. This video showcases the meticulous design process and the collaborative effort that went into building the pan-tilt system. It begins with an exploded view of the CAD model, highlighting the complex assembly of components, each designed with precision and durability in mind. Pay particular attention to the rotary seal assemblies, motor mounts, camera mounting system, and waterproofing (which successfully passed IP67 standards). These were my primary mechanical contributions to this project. These components are crucial for ensuring the system's high performance and reliability under varied environmental conditions.

Firmware Development
I acted as the primary software developer on the project, where I employed assembly language for Trinamic driver boards to breathe life into our mechanical design. The full firmware, along with detailed documentation, is available on Github.