Engineer, builder and permanent student. I work on safety-critical systems for the T-7A Red Hawk, study AI at Johns Hopkins, and spend the rest of my time building things across hardware, software and machine learning.
Scroll to climb
Where I've worked and studied
Wilshire Lane Capital
02Pilot card
About me
Every aircraft carries functions it simply cannot get wrong. Right now, my job is making sure they don't. As an Air Dominance Systems Engineer at Boeing, I work in Systems Engineering, Integration & Test on the T-7A Red Hawk, tracing how safety-critical functions can fail across more than 25 system threads and closing airworthiness gaps on the aircraft.
I'm most useful where disciplines meet. I trained in electrical and computer engineering at USC, and my work so far has spanned embedded hardware and firmware, model-based systems engineering and software. Moving between them has taught me to follow a problem from a single signal up to the behavior of a whole system, and to explain it clearly to the people who need to act on it.
I don't treat any of that as a fixed lane. I'm happiest when I'm learning something new, and at the moment that's artificial intelligence: I'm working through an M.S. at Johns Hopkins and building projects like Safe Autopilot to see how learned systems behave when the stakes are real. Whatever the next interesting problem turns out to be, in AI, software, hardware or a field I haven't touched yet, I want to be the one picking it up.
Loadout
Python
C/C++
MATLAB/Simulink
Verilog
JavaScript
React
SQL
MBSE
SysML / Cameo
Embedded systems
PCB prototyping
PyTorch
03Logbook
Work experience
BA
Air Dominance Systems Engineer
Feb 2026 – Present
The Boeing Company · T-7A, SEIT
Independently authored 4+ Safety Critical Function Thread Analyses within my first six months, identifying hazardous system interactions and failure pathways ahead of schedule.
Resolved airworthiness gaps through hazard analysis and cross-functional coordination on the T-7A.
Presented findings directly to government System Program Office stakeholders, who called it the most progress on the program's SCFTA effort in three years.
WL
Technical & Business Operations Intern
Mar 2024 – May 2025
Wilshire Lane Capital
Structured 1,200+ digital records with metadata in DealCloud, reducing lookup time by 60%.
Exceeded every internship KPI, earning a six-month extension.
04Sorties
Projects
ML / Autonomy2026
Safe Autopilot
A learned flight controller with a runtime safety monitor
A neural-network pilot flies a simulated Cessna 172, and a safety monitor hands control to a simple, known-safe PID autopilot the moment the aircraft leaves a safe envelope.
With the monitor watching, 0 of 180 deliberately broken flights crashed.
Details
I built the same safety pattern twice. In v1 the learned controller imitates the PID. In v2 it is trained with reinforcement learning (PPO). The monitor and fallback are identical in both, so I could compare the two learned pilots fairly.
Results
The RL pilot beat the imitation model in all five test conditions and beat the PID in every disturbed one. The PID is still the most precise in calm air.
I deliberately broke the learned controller's outputs with a stuck elevator offset and a stuck aileron. Flying alone, every learned controller crashed in the stuck-aileron test. With the monitor watching, 0 of 180 flights crashed.
Strong turbulence (severity 3+) crashes every controller, including the fallback, because none of them control engine power. I documented this as a known limit rather than hiding it.
What went wrong, and what I learned
My first RL reward taught the policy to crash on purpose, because ending the flight early scored better than continuing. I only caught it by reading the full training curve. I also found that my results changed between simulator versions, so I pinned the version and re-ran everything.
An interactive map of damaging bird-strike risk at 266 US commercial airports, treated as a system safety problem. A Bayesian hierarchical model estimates monthly damage rates with honest uncertainty, and a Mission Planner turns them into the chance of a damaging strike for a planned set of operations, classified on the MIL-STD-882E risk matrix.
Details
Built on 242,449 public FAA strike reports from 2010 to 2025. The negative binomial model pools information across airports, flyways and seasons, so only differences the data actually supports stand out on the map. You can play through the year to watch the spring and autumn migration waves.
Fitted on 2010 to 2022 and scored on 2023 to 2025, it beat a simpler baseline by about five standard errors on log score. Public data only, not for operational use.
Data pipelines and web apps built to production habits: scraping, local LLM cleanup, a database layer, tests and docs.
PostgreSQL
SQLAlchemy
Flask
pytest
GitHub Actions
HardwareJan – May 2024
Terra Tech Plant Health Sensor
Lead hardware / electrical engineer
USC senior capstone: a battery-powered platform integrating moisture, humidity and NPK soil sensors. Power management extended system efficiency by 35%, validated to better than 90% accuracy.
Embedded firmware
PCB design
I²C / SPI / UART
Course ventureJan – May 2024
REFINE.AI
Hardware lead
A startup concept from USC venture coursework. I designed the hardware architecture and prototype concept and built the bill-of-materials analysis for the team's pitch.
Hardware architecture
Bill of materials
Web / Pixel art2026
Flight Log
Design & engineering
The site you're reading. A pixel-art jet drawn on canvas flies a flight plan keyed to your scroll, with a working HUD, waypoint display, day-to-night sky and a mode where you fly it yourself.
Canvas 2D
JavaScript
Pixel art
05Training record
Education
JH
Johns Hopkins University
Aug 2026 – Present
M.S. in Artificial Intelligence
Whiting School of Engineering, Engineering for Professionals program.
SC
University of Southern California
Aug 2021 – Dec 2025
B.S. in Electrical & Computer Engineering
Viterbi School of Engineering, computer engineering emphasis. Coursework in embedded systems design, computer systems organization, robotics, linear systems and circuits, and electromagnetics.
EC
Eton College
2016 – 2021
A Levels: Mathematics, Physics, Economics
06Comms
Contact
Systems engineering, safety-critical software, or AI in aerospace. If it's a hard problem, I'd like to hear about it.
Wilshire Lane