About
Systems Administrator at Denison University and applied machine learning researcher. Originally from Nepal; currently based in Granville, Ohio.
Most of my research time goes into small, careful machine learning systems on real datasets — image captioning with CNN encoders and LSTM decoders, climate pattern classification with SOM clustering and CNN classifiers, RNN vs. CNN comparisons on clinical tabular data, and text classification with both classical and neural pipelines. The thread across all of it is the same: take a domain seriously, evaluate honestly with the right metrics, and write down enough that someone else can reproduce the result.
More recently I've been building MedCompress, a tool that condenses long medical documents into structured, faithful summaries that downstream NLP pipelines and clinical decision-support workflows can actually use. The motivating question is simple — how do you keep clinical signal intact through aggressive compression — and the underlying engineering work has pushed me deeper into modern NLP, evaluation design, and how ML systems get adopted in places that can't tolerate ambiguity.
I studied Computer Science at Beloit College, graduating Summa Cum Laude with Phi Beta Kappa honors and a 3.87 GPA. Coursework spanned Machine Learning, Data Mining, Software Engineering, Database Systems, Algorithm Design & Analysis, Computer Architecture, Threads & OS, Mathematical Statistics, Real Analysis, and Differential Equations.
Alongside coursework I worked as a Systems Programmer Intern at Beloit College, designing Python and PowerShell automation against Active Directory for the full student account lifecycle, and as an IS Network Services Professional Intern at Wisconsin ECB, supporting Cisco network upgrades and building Power Automate reporting flows.
I care about ML work that takes the underlying domain seriously instead of treating it as a generic tabular task, and about systems that respect the people who use them — clear evaluation, honest reporting, and documentation that holds up after the original author has moved on. Outside of research I cook Nepali food, hike whenever the Ohio weather allows it, and read about how new technology lands inside institutions without breaking them.