Full stack Android app planning and development utilizing multiple AI models

● Overhauled power management firmware for Zephyr-based cameras, refining charging algorithm and updating proprietary USB-C Linux drivers to optimize battery charging, prevent accidental BMS learning cycle triggers, and improve overall efficiency.● Quickly acquired necessary proficiency in C, kernel C, and Linux kernel interaction using Device Tree files to enable seamless code integration into existing codebases and fast comprehension of existing code.● Devised custom algorithms for wind noise reduction to enhance camera audio quality, involving rapid learning of technical skills such as audio windowing, windowed processing techniques, audio re-stitching, and real-time GStreamer pipeline creation using custom elements.● Proposed 5 algorithm proposals in Python that ran with minimal memory and CPU usage, emphasizing kick sounds and game sounds while reducing wind noise.

● Conducted extensive verification testing and documentation on pre-release 16 – 50V input, 16-60V output DC – DC power converter line.● Trimmed switching timings to improve efficiency across the line while simultaneously preventing board failures due to tolerance differences.● Optimized features such as overvoltage turnoff for 20 C, 100 C, and -40 C temperatures, turn-on ramp-up and overshoot, and switching overshoot.● Inspected and corrected localized board issues which prevented further optimization, such as excessive noise on signal lines.● Wrote design verification (DV) document cataloging all findings and actions.● Developed and proposed solutions to failures observed through testing functionalities

● Hired as first hardware engineering intern at growth stage, Greentech company in a self-directed role.● Supplemented two-person hardware team by self developing tasks and learning additional skills necessary to perform tasks.● Created simulations emulating expected worst-case product current draw using Python and C++ to aid battery sourcing decisions.● Self studied Python to create the script that interacted with the electronic load and ran current draw simulations.● Installed and tested prototype devices in municipal trash cans in Denmark and San Francisco.