Projects & Research
Embedded systems, edge ML, and reinforcement learning for antenna design.
Current Research
RL Tuned Dipole
A lightweight RL environment wrapper around NEC2 for automated dipole antenna resonant frequency optimization.
antenna-designgymnasiumnec2pythonreinforcement-learning
Full writeup & results
Scaled Dot-Product Attention
Fixed-point scaled dot-product attention in Verilog: two shared MACs, LUT-based softmax, Q8.8 datapath, Verilog testbench and a Python reference model.
attentionfixed-pointhdlrtlscaled-dot-product-attentionsoftmaxverilog
Full writeup & results Paper Replication
Deep Reinforcement Learning for Dynamic Multichannel Access
PyTorch replication of Wang et al. (2018) "Deep Reinforcement Learning for Dynamic Multichannel Access." DQN vs. the analytical optimal policy on the round-robin switching case
deep-q-networkdynamic-spectrum-accesspaper-replicationpomdppytorchreinforcement-learningwireless-networks
Full writeup & results Publication
Fall Detection System (IoMT)
Wearable fall-detection device with double-layer LSTM (97.8% accuracy), companion to published research
embedded-mlfall-detectionhealth-monitoringiomtlstmpcb-designrp2040wearable
Full writeup & results
Star Tracker Simulator
Synthetic star-image simulator and visualization layer for spacecraft attitude determination
adcsattitude-determinationcentroidinghipparcospythonspacecraftsstar-tracker
Full writeup & results
IoT Streaming Firmware
C-based IoT sensor data streaming firmware
cesp32iotmqtt
Full writeup & results