Bernardo
Lozano
Electrical engineer working on machine learning for imaging and predictive maintenance — signal processing, time-series models, and the benchmarks that decide which one actually ships.
a tuned LSTM
single-frame baseline
bilinear interpolation
Fabricated working integrated circuits on a silicon wafer in Notre Dame's NDNF cleanroom — a semester-long process spanning photolithography, oxidation, diffusion, and metallization, ending in electrical test of the finished dies.
Custom KiCad PCB around a Raspberry Pi Pico 2 with LoRa and GNSS modules and USB-C/LiPo power-path management; firmware evolved from a superloop into a layered FreeRTOS design with ISR-driven GPS/LoRa I/O.
Led a 12-person Data Club team building a job board for students — scraping pipeline, SQLite store, and a scikit-learn NLP matcher pairing resumes against 680+ scraped postings.
3-axis Helmholtz cage driver generating precise ±125 µT fields with PWM-driven H-bridges and PID control in C++, cancelling Earth's magnetic field to ±0.1 µT to simulate low-Earth-orbit conditions for satellite hardware.
C++ team project loading county-level energy data, computing per-capita consumption, and rendering an interactive 3D scatter-plot map that surfaces usage patterns across US regions.