I like the part where it stops being theory
Every project on this site hit the same wall at some point: the idea worked on paper, and then the hardware, the network or the free-tier quota disagreed. That gap is the part I find genuinely interesting.
On the MolKet benchmark it was IBM Quantum's monthly measurement allocation — VQE was fully built, ansatz and Hamiltonian and optimizer, and simply could not be run. On Echoes From the Shadows it was a 5V supply and a mobile hotspot deciding what video resolution I was allowed to have. Neither is in the textbook.
Leading a project is most of the work
I was project lead on the MolKet benchmark: a multi-month comparison of four ground-state solvers, run with Taha Selim as client and Marten Teitsma as coach.
That meant setting up the team's collaboration framework and issue tracking on day one, and keeping the project plan current the whole way through — tracking progress and risk so we hit our milestones on schedule. The algorithms were the visible output. Getting a research codebase across the finish line on a deadline was the actual job.
I also wrote two of the four solvers myself — VQE and Exact Diagonalization — and contributed the first iteration of SQD.
Hardware and software, in the same repository
Echoes From the Shadows started as a question about whether I could get an ESP32-CAM to stream somewhere useful, and turned into a dashboard that discovers hosts on the network, saves them, pulls multiple live streams at once and records audio through an attached microphone.
The camera module itself was the easy half — the documentation is excellent. The hard half was everything around it: connection handling, stream latency, and building an interface that stays usable when a camera drops off the network.
Currently
Working through a quantum computing minor, which is where the Qiskit habit came from, and building small developer tools when something annoys me often enough to be worth fixing.
