P / {{ p.no }}
{{ p.title }}
{{ p.desc }}
I'm a biomedical engineer who took one too many neuroscience courses and never looked back. These days I think of myself as a computational neuroscientist in progress, interested in how the brain organizes and refines the things it learns to do. I like working at the level where theory and data have to talk to each other, and most of my questions, one way or another, find their way back to the motor system. How it learns, how it adapts, and how it does any of it at all.
My thesis at Brown focused on reconstructing SEEG signals corrupted by cortical stimulation artifacts, using a convolutional autoencoder paired with an LSTM to recover connectivity structure that would otherwise be lost. It was a different kind of problem, but it taught me how to build models that have to hold up under messy, real-world constraints, and that's not nothing. Outside of research, I follow Formula 1, love exploring unique food spots, and stay sane through running and cycling.
Built a text-to-CAD platform using LLMs and RAG to generate parametric CAD models from natural language, cutting design iteration time by 70%. Programmed automated liquid-handling systems and prototyped lab instrumentation for synthetic-biology workflows.
Developed a CAE-LSTM model reconstructing SEEG signals corrupted by stimulation artifacts, improving fidelity by >40% while preserving phase and temporal dynamics. Processed 500+ hours of human SEEG and contributed to closed-loop pipelines for Class III neuromodulation systems.
Fabricated PVDF piezoelectric motion sensors in an ISO-class cleanroom and integrated them into wearable fall-detection prototypes capturing joint kinematics and muscle activity. Bench-tested with 15+ subjects for neural-to-prosthetic interface applications.
Prototyped a pediatric EEG headset, electrode layout, PCB routing, enclosure. Developed EEG preprocessing pipelines (ICA, artifact rejection) and ML models to detect electrophysiological markers of autism in large-scale pediatric datasets.
Thesis: reconstructing SEEG recordings during cortical stimulation using a convolutional autoencoder–LSTM architecture.
Thesis: multiple sclerosis lesion segmentation in 3D MRI.
{{ p.desc }}