Zachery Pipes

Rolla, MO

• Directed three engineering sub-teams (9+ members) spanning flight simulation, developer tooling, and 3D terrain mapping
• Led migration of the team simulator from legacy AirSim to Project AirSim, containerizing the Unreal Engine and ArduPilot SITL stack (Podman/WSL2) and restructuring it as a shared submodule serving the design team's competition codebases
• Cut new-member setup from a multi-hour source build to a single pull by publishing prebuilt environment and SITL images to GitHub Container Registry behind a one-command launcher
• Structured sub-team work around end-to-end feature ownership rather than by layer, giving every member hands-on work with the team's core technologies
• Led onboarding for 5+ new members, training them in Python, Git submodule workflows, and containerized development environments

Rolla, MO

• Developed C++ software for a UAV platform, contributing through code review and testing
• Collaborated with 10+ teammates in a student-led UAV startup advancing next-generation UAV technology under active NDA
• Modeled product components in Blender and produced marketing renders used in the company website and pitch decks
• Evaluated candidate UAV components against Blue UAS sourcing requirements, weighing compliance against performance constraints

Columbia, MO

• Processed and shipped packages to destinations across the United States
• Trained new employees on store procedures and best practices
• Provided excellent customer service through answering questions, resolving issues, and providing solutions tailored to customer needs,

C++, Python, CUDA, PyBind11

• Implemented 26 CUDA kernels accelerating 5 probability distributions and vector utilities, with a streaming executor that splits large inputs across 4 concurrent CUDA streams
• Built the pybind11/NumPy wrapper layer between C++ and Python, releasing the GIL during computation and managing GPU memory through self-freeing buffer objects that are allocated once and reused across calls, replacing manual cleanup on six separate error paths
• Achieved up to 57x speedup over SciPy with the C++ CPU backend alone, with CUDA kernels adding a further 2.5x over the CPU path on large inputs
• Automated cross-platform wheel builds for Linux, macOS (x86_64/arm), and Windows across Python 3.10--3.14, with trusted-publishing releases to PyPI and a CI check that catches type-stub drift against a CUDA build

GPA: 3.7

Intro to Networking, Databases and File Structures, Algorithms, Intro to OS, Theory of CS, Data Structures, Intro to C++