What I Do
A few things I do, and can show you have worked. Fewer, but done well.
Semiconductor Imaging & Metrology
Building a coherent imaging-and-measurement pipeline at NIST that carries work from raw data to a result people can actually use — a few projects deliberately woven into one usable tool.
Scientific Imaging & CV
Tomography, segmentation, and 3D reconstruction across microscopy, neutron, and SEM data — including synthetic data for model training when real data is scarce.
ML / DL with an eye on interpretability
Models that have to earn their keep in science — classification, forecasting, and segmentation where being able to explain the answer matters as much as being right.
Data & Scientific Software
The unglamorous, reliable layer: containerized workflows (Docker, WIPP), HPC scheduling (SLURM), and pipeline design that makes research reproducible and painless to rerun.
Data Science & Analysis
From messy, real-world data to answers that survive scrutiny — statistics, forecasting, and careful analysis, in science and far beyond it.
Critical Analysis & Debugging
When something behaves strangely, I find out why — once catching a fundamental flaw in a colleague's method before a major conference, then improving the work and joining the paper.
About Me
I'm genuinely curious about a lot of fields — but what keeps me grounded is wanting to understand how things work, and a taste for doing it well.
I started in metallurgy and materials at IIT-Kharagpur, earned a PhD in polymer science in Akron, and somewhere on the way fell for what machine learning and imaging could do for science. That's taken me through startups, academia, and government research — and taught me to pick up new domains fast, from HPC and scientific software to bench-scale formulation.
Today I do applied AI and scientific image processing at NIST. Along the way I keep an eye on where intelligence seems to be heading — and lately, living systems have a hold on my imagination.
Say Hello
I enjoy hearing from people doing interesting work. Especially if you're wrestling with scientific imaging or messy data, building research software people might actually use, or thinking about living systems and where intelligence is heading — I'd genuinely like to compare notes.
Get In TouchRight Now, I'm Curious About
The kind of things that pull me in lately. Expect this to change — I like keeping it fresh.
Imaging the Unseen
Tomographic reconstruction and scientific image processing — pulling real signal out of noisy, complex microscopy data.
Reproducible Scientific Software
The quieter side of the work: containerized pipelines, HPC orchestration, and tooling built to be rerun — so a result someone depends on can be produced again, on demand.
Living Systems & New Intelligence
How intelligence works in biology — and what happens when we build machines that borrow the same tricks.