SPC Fall 2024 Founder Fellowship Cohort

Software + Firmware ConsultingPast client projects include: * Rust development for hardware-in-loop testing application * Custom camera software development and hardware bring-up on Qualcomm Snapdragon platform * Micropython application port to C on an RP2040 MCU * STM32 prototype firmware development

Led Ai Pin's system software and firmware development from prototype through mass production. My team's responsibilities included developing a highly customized Qualcomm Android BSP, performance & power optimization, frameworks, GPU acceleration, MCU firmware development, and manufacturing test support.Collaborated with founders, EE, ME, Ops, and Design to define product architecture and select componentsGrew system software + firmware org from 1 (me) to 30+ FTE, plus additional consultantsDeveloped first- and second-line management team, processes, and succession plan for my orgDelivered foundational product features and factory test software in C, C++, Java, and PythonScoped work and hired early talent in other key software domains (eg, test automation, wireless), then spun those core components out to new teams and leaders once they hit critical massPartnered with program management on schedules, roadmaps, scoping, and deliverablesDeveloped business relationships and onboarded vendors through legal and finance processes

Prototyped mixed-reality passthrough camera system softwarePrototyped eye-tracking camera sync and illumination firmware and software

Managed team of 2 (start) --> 7 (end) engineers developing camera software support for NVIDIA Jetson platform's Android and Linux BSP (Board Support Package) releases. The project-management side of my role was driving a regular release cadence of bugfixes and feature support within the broader camera team as a matrix org structure, and communicating regular progress and risks to cross-functional and leadership stakeholders.The people-management side of my role included providing regular coaching, performance management, and career growth support to my direct reports. My team's work included delivering CUDA-accelerated image processing features, which required collaboration across engineers with expertise in drivers, CUDA implementation, algorithms, and image quality tuning.A personal career highlight was coordinating 25+ cross-functional engineers to deliver successful camera software subsystem chip bring-up for the Jetson AGX Xavier platform.

Eyefluence built eye-interaction hardware and software solutions for head-mounted displays (ie, AR/VR headsets). The company was ultimately acquired by Google ~Oct 2016.I served as the embedded SW lead and provided essential cross-functional glue between hardware, algorithms, UI/UX software, and the executive team to help bring our prototypes and demos together. Usually on extremely tight timelines ;)Highlights: - Managed 2 engineers and 1 contractor - Developed Linux drivers and V4L2 (Video4Linux) integration for MIPI CSI global shutter cameras - Developed Linux I2C drivers for SW control of an FPGA for custom camera + illumination sync design - Reviewed schematics with the hardware team - First-in-line for HW bringup and debugging post-assembly - Worked closely with HW to triage issues and identify reworks and bugfixes - Architected SW layer for embedded vision processing and client event consumption across Android JNI - Supported algorithm engineers in porting Matlab algorithms to embedded C for real-time implementation - Developed Android SDK for eye-interaction and supported customers in adoption and prototyping - Served as go-to person for troubleshooting and triage for basically everything, including supporting C-level team with hotfixes and last-minute additions to high-priority demos

Translate Abroad was a small (<5 person) startup developing the Waygo mobile app. The app performed real-time offline computer vision and translation to overlay English text on foreign-language menus and signs.I led the effort to port a pre-existing iOS app to Android. This included an extensive refactoring of a monolithic Objective-C file into a shared native C++ SDK that decoupled computer vision algorithms from the UI/UX layer, and Android UI/UX development. Project successfully completed and shipped on the Google Play Store.Instituted testing, code review, and bug tracking workflows in the team. Built library of ground-truth data for automated algorithm regression testing, and created tools for annotating large quantities of test data.Work also included some reference computer vision algorithm development in Matlab.

Android camera driver development in C, C++, and Java. Work spanned Android framework, HAL, and Linux kernel. Supported product launches of HTC One X, Motorola Xoom, Motorola Atrix, and more.Led re-architecture of camera stack that delivered 10x improvement in performance KPIs by eliminating redundant layers and unnecessary overhead.Developed GPU-accelerated computer vision and computational photography algorithms, including optimization of an HDR image fusion algorithm that reduced total processing time on an image batch from 30s to 3s.

Developed camera platform-level features in Android on the NVIDIA Tegra platform. Debugged critical customer issues in support of the Motorola Atrix program. Languages and stack layers primarily consisted of C and C++ in the Linux Kernel and proprietary driver layer respectively, with some Java to integrate with the Android Framework.

Perl scripts for silicon validation - round 2! This year focused on extending the web interface to the validation scripting from the prior year.

Wrote Perl scripts for silicon validation. Automated GPU performance regression testing and generated user-friendly HTML reports comparing performance across test runs. Used during pre-tapeout HW development simulation and emulation cycles.

Wrote Perl scripts to generate automated reports from C/C++ static analysis tools to improve modem firmware code quality