About
Currently seeking neural signal processing / BCI engineering roles — open to remote or the NC Research Triangle.
I'm a neural signal processing and BCI engineer. My largest project is a first-author study on the neural sub-processes that link speech perception and production — a PyTorch tensor decomposition (sliceTCA) over intracranial-EEG recordings from 31 patients — and most of my PyTorch and signal-processing experience comes from building and scaling it. Most recently, in the Cogan Lab at Duke University's Department of Neurology, I also worked with high-density μECoG arrays (up to 1024 channels) recorded intraoperatively at Duke University Medical Center — a collaboration spanning Duke Neurology, Neurosurgery, and the Viventi Lab's hardware engineering team, under FDA Investigational Device Exemption and Duke IRB oversight.
I built and maintained IEEG_Pipelines, the lab's open-source iEEG/ECoG analysis stack (BIDS conversion, PHI removal for NWB/DANDI compliance, GPU-accelerated signal processing). Every change ships through code review on pull requests, tested across three OSes via GitHub Actions before merging; 1,243 of its 1,423 commits are mine, version-controlled in Git alongside the co-maintainer I trained and other contributors. Fittingly, this site's color palette and type are modeled directly on its documentation theme.
This site is itself hand-built — no theme, no framework — including the search box above and the light/dark toggle; see the colophon for how it's put together.
I have a real passion for brain-computer interfaces, medical data analysis, and research more broadly — and I'm open to new opportunities in that space.
Experience
- Scientific Data Manager I — Duke Department of Neurology (Cogan Lab). Led data infrastructure for research under FDA IDE and Duke IRB oversight; maintained BIDS-standard systems with PHI removal for NWB/DANDI compliance; authored and version-controlled analysis pipelines; led the lab's migration from MATLAB/Windows to a Python/Linux stack, managing a live clinical pipeline cutover across the whole lab.
- Research Technician II — Duke Department of Neurology (Cogan Lab). Converted lab data systems to the BIDS standard; built new ECoG/iEEG analysis pipelines; supported intracranial recording research through that same cross-functional collaboration — clinical, surgical, and hardware engineering — described above.
- Lab Technician — Cornell Psychology Department. Built an AWS-deployed fMRI preprocessing server (BIDS format), integrating FSL, AFNI, FreeSurfer, and ME-ICA via Python Nipype — see CMRIF Preprocess.
- Research Assistant — Cornell Psychology Department. Supported an fMRI study of Locus Coeruleus activity; identified a study-design flaw that led to a protocol change.
Education
- Duke University, Pratt School of Engineering — Master of Engineering, Biomedical Engineering; Medical Device Design Certificate.
- Cornell University — B.S., Bioengineering.
- George Washington University — one year of undergraduate study before transferring to Cornell.
Publications & presentations
First-author manuscript (in preparation):
- Earle-Richardson, A. M., Duraivel, K., Southwell, D., Sinha, S., Vestal, M., Grant, G., Zafar, M., & Cogan, G. B. Neural sub-processes linking speech perception and production. Manuscript in preparation — presented as posters at the Society for Neuroscience (2022–2024), Duke TBS (2022–2023), and AAAS (2025). Read more →
Selected co-authored presentations:
- Sexton, D. P., Earle-Richardson, A. M., Southwell, D. G., Vestal, M., & Cogan, G. B. Decoding verbal working memory load from intracranial high gamma activity. AAAS, 2025.
- Zhang, J., Earle-Richardson, A. M., Southwell, D., Egner, T., & Cogan, G. B. Intracranial EEG correlates of concurrent demands on cognitive stability and flexibility. Society for Neuroscience, 2024; Cognitive Neuroscience Society, 2024.
Skills
Day to day the split is: PyTorch and GPU work on the tensor-decomposition side (sliceTCA), classical ML (PCA-LDA, scikit-learn) for the decoders, and the high-gamma signal processing and data infrastructure underneath both.
Programming
Python (PyTorch, MNE-Python, NumPy, SciPy, scikit-learn, Nipype), C/C++, MATLAB, Bash, CUDA, HTML/Liquid, JavaScript
Signal & data
iEEG/ECoG/EEG, fMRI/DTI, high-gamma extraction, GPU-accelerated tensor decomposition (sliceTCA), time-resolved decoding, permutation statistics, BIDS, NWB/DANDI
Platforms
GitHub CI/CD, GitHub Pages, AWS, Linux, ReadTheDocs, Duke Compute Cluster
Regulatory
Design Controls (21 CFR 820.30), ISO 14971 risk management, predicate-based 510(k) strategy