At the Center for Computational Mathematics, I architect and implement software systems that support collaborative scientific research: data sharing, visualization, and process automation. Main domain is currently computational neuroscience/electrophysiology (spike sorting).

Projects include:

• kachery (a scalable system for cloud storage of large scientific data files)

• FigURL (a framework for developing richly interactive, browser-based visualizations of datasets with shareable permalinks)

• Visualization modules for electrophyisology research, 3D vector/scalar field viewer, and real-time process monitoring

• Scientific data pipelines on cluster environments

I also sometimes assist with an outside lab that applies inverse Reinforcement Learning to problems in healthcare (ACM paper 2021: https://dl.acm.org/doi/10.1145/3459930.3469536) and I'm an organizer and frequent presenter in an internal training series on software engineering best practices.

A senior software designer and full-stack developer for four-person team working primarily in C#, SQL, and JavaScript. Also extensive client management work, from requirements gathering and needs assessment, to business process redesign, to ongoing communications during the development process.

Designed and implemented searchable online database of human-reviewed published SEC filings (Securities Enforcement Empirical Database) for quantitative legal research (C#).

Sole designer and principal developer of system, including tool to enable independent two-party review of documents per professor's research methodology;

Managed assistant developer during project's transition to maintenance.

As part of five-person team, developed online course bidding and management system which assigns courses to 1400 students twice annually using a blind second-price multi-winner auction with transferrable bids. (C#, Vue) (This system is only accessible to current NYU Law students.)

Wrote data integrity, decision support, and data presentation reports for Alumni Relations and Financial Services, saving over 40 hours of work per week.

Independent research for thesis on applying machine learning methods to predicting harmfulness of protein mutations. Heavy use of Python and standard ML/scientific packages (SciKit-Learn, NumPy, PyTorch, XGBoost, and others). Rewrote existing codebase to be more modular so that I could run comparative experiments using different types of classifier.