Shantanu Parmar

I've always been drawn to what lies beyond √2GM/R, and my work is about understanding a little more of it. Astrophysics is at the heart of what I do, with machine learning, imaging, and electric propulsion as the tools that help carry it further.

Visiting student at Cornell University, Ithaca, New York.

Imaging and computer vision

Reading the sky from the ground

Event-based star tracking, generative restoration of low-SNR astronomical images, and object detection benchmarking.

Electric propulsion

Predicting how plasma engines behave

Machine-learning performance prediction for Hall effect thrusters, and plasma plume simulation for electrospray thrusters.

Astrophysics and gravitational waves

Listening for ripples in spacetime

Pulsar timing, white dwarf binaries, and transient noise classification in LIGO data.

Scroll to pull back from Earth. Skip the intro

Recent and ongoing work

White dwarf period derivative measurement
16th LISA Symposium, poster

PDot Measurement for Detached White Dwarf Binaries as Future LISA Sources

Mar 2026 – present, supervised by Dr. Warren Brown, Smithsonian Institution and CfA at Harvard University

Galactic double white dwarf binaries are a prolific LISA source class, though most detached binaries appear as mono-frequency sources. This work presents optical time-series spectroscopy for known LISA white dwarf binaries, fitting the radial-velocity phase across epochs to demonstrate that orbital-frequency-change measurements are possible for non-eclipsing systems, enabling physical understanding of these future multi-messenger laboratories.

ML performance prediction for Hall effect thrusters
Journal of Electric Propulsion, under review

Machine Learning Techniques for Performance Prediction in Low-Power Hall Effect Thrusters Operating on Monoatomic and Molecular Propellants

Nov 2025 – Jul 2026, supervised by Dr. Dan Lev, HPEPL, Georgia Institute of Technology

The shift toward low-power Hall effect thrusters and alternative molecular propellants requires predictive models that go beyond traditional monatomic empirical scaling laws. This work develops a propellant-agnostic machine learning framework to predict thrust, beam current, and discharge current across varied propellants (Xe, Kr, Ar, CO₂, N₂) in low-power regimes.

Also building

Independent projects and tools.

Mobil-Telesco

An astrophotography aid.

HETCalc

Parameter calculator for solenoids in Hall effect thrusters.

Open to collaboration and conversation

Reach out about research, black holes, electronics, or anything in between.