Amogh D. Prasanna
Graduate researcher at the University of Maryland, College Park, working with Prof. Shabnam Jabeen and, at Fermilab's SQMS Center, Dr. Tanay Roy.
Nature isn't classical, dammit, and if you want to make a simulation of nature, you'd better make it quantum mechanical.
I'm interested in quantum error correction, quantum simulation and quantum complexity: how to keep quantum information intact, what a quantum computer can simulate that a classical one can't, and how to prove where that line sits.
So far I've met those questions on real hardware: compiling qudit unitaries into native gates on transmons, decoding topological codes, and training quantum generative models that have to survive real-hardware noise. On a noisy device the budget is two-qubit gates, not depth; on a qudit, leakage and anharmonicity belong in the objective, not in a footnote.
With Dr. Tanay Roy at Fermilab's SQMS Center, I'm training models that compile arbitrary SU(d) unitaries into native gate sequences for superconducting transmon qudits, built on the universal Jaynes-Cummings interaction in cQED. Leakage and anharmonicity are part of the objective from the start, not patched on afterwards.
Read more: Can a learned compiler beat analytic decompositions on a qudit?There is no first-principles calculation of hadronization, so event generators like Pythia fit a phenomenological Lund string model instead. I train quantum circuit Born machines on Pythia 8 momentum distributions and check what survives once IBM's noise model is switched on.
Read more: Can a quantum circuit learn how quarks become hadrons?32native ECR gates in the best circuit under FakeBrisbane noise
write-up in progressI'm contributing to an open-source Python and C++ toolkit for simulating and decoding topological codes, built around Stim and PyMatching. My part so far is algorithm frameworks, pull request review and benchmark analysis.
Read more: Decoding topological codes.Quantum machine learning for hadronization (working title).
Python · Qt · Claude API
A study companion that floats over the screen. Ask about whatever you're reading or watching, and markers fly to the symbols the answer is about. The hard part was landing them on the right pixels, so it resizes each screenshot itself and snaps every point onto the OCR line the answer names. Every question goes into a dated journal of doubts.
YC × Cactus × DeepMind hackathon, team project
A team-built, on-device voice assistant for HVAC technicians. I wrote the search behind it, for when the manual's answer doesn't fix the unit. My ranker lets a certified technician's deeply nested forum reply outrank the official support page by weighing author expertise, reply depth and "this fixed mine" replies against source authority. It is unit-tested but not yet wired into the app.
HVAC Copilot · Python, on-device Gemma
A course project that takes a plain inventory app through misuse cases and STRIDE. 29 threats became 38 security requirements, and the code cites 27 of them by ID. The clearest change was moving roles out of client-held tokens and into server-side sessions reloaded on every request.
Threat-Model-Dunder-Mifflin · UMD ENPM680 · Go, Fiber, PostgreSQL
TAK's stock CoT injector only scatters static markers. I added a simulated GPS unit that drives line, circle and square paths at a set speed through ATAK's internal dispatcher, so the map treats it like a real GPS-reported friendly unit and draws its trace.
ATAK-CDCL · Java, Android TAK SDK
Before graduate school I spent two and a half years shipping production code, which is why my research code is tested and reproducible.
Updated October 2026
One qubit, simulated in your browser. Apply gates, measure it, or turn up the noise and watch the state lose its purity: the Bloch vector shrinks toward the centre and the sphere comes apart.
There is also a real shared qubit on my GitHub profile: everyone who visits acts on the same state.Apply H to it
Outside the lab I paint, take photographs, and box.
Capturing light and moments: the same physics that governs photons, frozen in a frame.
Creating art with the same curiosity that drives research: exploring form, color, and expression on canvas.
Boxing, and not boxing, until someone opens the gym door.