- Designed a multi-cycle RV32IM processor in Verilog around a fetch, decode, execute, memory, writeback state machine.
- Wrote a self-checking testbench and an 88-case assembly test plan covering the RV32I base ISA and M extension, including divide-by-zero and overflow edge cases.
- Automated an assembly to hex to simulation regression flow with timeout detection and per-test pass/fail reporting.
- Built a 640x480 HDMI controller with TMDS encoding and a CPU-writable frame buffer, using a pixel clock enable to avoid clock domain crossing. Demoed Snake running on hardware.
Projects
150 µs end to end, PHY to parsed payload
- Wrote bare-metal C on the Zynq ARM core to pass Ethernet frames over AXI to a SystemVerilog byte-stream parser.
- Measured end-to-end latency on hardware while extracting payloads from 80-byte frames.
- Built a UVM environment with an agent (sequencer, driver, monitor) and a scoreboard to verify frame handling.
- Designed a two-layer board in KiCad around six 18650 Li-ion cells wired in parallel for capacity, with an IP5306 managing charging and protection from USB-C.
- Boosted the 3.0 to 4.2 V pack to 15 V with an MT3608, placing the switch, inductor, Schottky diode and capacitors tightly together to keep the high-current switching loop small.
- Brought up the board after finding that the capacitor footprints didn't match the parts I received, rewiring them with jumper wires instead of waiting on a new board.
- Tested on a phone, which charged but reported slow charging. Next revision: a USB Power Delivery controller so the phone can negotiate fast charging, and footprints checked against the datasheets of the exact parts ordered.
Experience
-
May 2026 to now
Machine learning researcher, BOLD Lab
University of Colorado Boulder
Predicted magnetic core loss at 4.78% error across 4 unseen materials with BelNet and DeepONet models in PyTorch, using excitation waveform, frequency and temperature.
-
May to Aug 2025
AWS engineer intern, Exterro
Coimbatore, India
Deployed EC2 instances and S3 buckets, managed over SSH on Linux, to store and process document datasets.
-
May to Aug 2024
ML engineer intern, Exterro
Frisco, TX
Cut document review time by 60% with a TensorFlow classifier that sorts files by type, at 91% accuracy across 50K+ documents.
Education
B.S. Electrical and Computer Engineering, minor in Quantum Engineering
University of Colorado Boulder, expected May 2028
Digital Logic, Programming Digital Systems, Microelectronics, Programming Embedded Systems, Electronics Design Lab, Advanced Circuit Analysis