About
Aaron-James Santa Cruz
Experience
- Mechanical EngineerStealth-stage startup
End-to-end physics-based digital twin of a precision motion platform; sub-25-micron accuracy under load.
Jan 2026 – present - Undergraduate ResearcherMIT LEAP Group
Real-time machining-chatter detector: 97.1% on unseen noisy cuts, where a trivial baseline sits at 58%.
Sep 2025 – present - Mechanical Engineering MemberMIT Arcturus
Modular, instrumented tow-tank hull and multiphase CFD at up to 75% lower compute cost.
Sep 2025 – present - Lead Mechanical Engineer (Intern)Brigham and Women's Hospital
Mechanical design lead on an autonomous tissue-targeting insertion device; a loading mechanism for 5× sample capacity on standard labware.
Mar 2026 – present - Mechatronics InternQT Sense
Filter changer redesigned from $350 to $15; enclosure lead time cut from 8 to 2.5 weeks; sub-microliter liquid handling on a 10 nm piezo stage.
May 2026 – present - Mechanical / Electrical InternGeolabs, Inc.
A 360° survey drone (2.5 G, about 55 mph) validated in FEA, a four-layer STM32 FOC motor controller, and drone photogrammetry flood models.
May 2023 – present
Education
- Massachusetts Institute of TechnologyBS Mechanical Engineering + EECS2025 – present
- Leeward Community College2024 – 2025
I’m a Mechanical Engineering and EECS undergraduate at MIT (Class of 2029). I work across the full stack of a physical system: mechanical design and FEA, motor-control electronics, and the machine learning that turns noisy sensor data into decisions.
My work spans a precision-motion digital twin for a stealth-stage startup, a real-time machining-chatter detector in an MIT lab, and an autonomous tissue-targeting device at Brigham and Women’s Hospital. Most of it is about de-risking hardware before it’s built.
How I work
Mechanical & simulation. Static, modal, and harmonic FEA; multiphase CFD and thermal; parametric sweeps and mesh-independence studies, each run to settle a design decision. I’ve turned a failed 100 GB solve into a 34-second run so a team could iterate in a loop.
Electronics & controls. Motor-control hardware (field-oriented control, back-EMF sensing, thermal management) on multilayer boards, from a 6S FOC ESC to piezo motion at 10 nm resolution.
Perception & ML. Signal-only, CPU-only detectors that survive the gap between clean lab data and a real machine, plus computer-vision pipelines that replace slow manual work.