A Real Quantum Hackathon

Challenge Specifications & Starter-Kit Baselines

Harmoniqs × Pasqal × Microsoft · July 29, 2026 · Microsoft Garage, NYC

• Event page: harmoniqs.ai/hackathon
• Questions during the day: your Discord team channel · discord.gg/VJnXXeJda
• The three challenges live as issues: Challenge 01 · Challenge 02 · Challenge 03

The model

All three challenges use the same system: a register of $N$ atoms evolving under the analog-mode Rydberg Hamiltonian, driven by a single global pulse,

$$ \frac{H(t)}{\hbar} \ =\ \frac{\Omega(t)}{2}\sum_i \sigma_x^{(i)} \ -\ \delta(t)\sum_i n_i \ +\ \sum_{i \lt j}\frac{C_6}{\hbar, r_{ij}^6} n_i n_j , $$

where $\Omega(t)$ is the Rabi frequency and $\delta(t)$ the detuning of the global drive (shared by every atom), $n_i = |r\rangle\langle r|i$ counts the Rydberg excitation of atom $i$, and $r{ij}$ is fixed by the register layout. Your control variables are the waveforms $\Omega(t)$, $\delta(t)$ and the atom positions.

The blockade radius $R_b$ is the spacing at which the interaction equals the drive,

$$ \frac{C_6}{R_b^6} = \hbar\Omega \quad\Longrightarrow\quad R_b = \left(\frac{C_6}{\hbar\Omega}\right)^{1/6}. $$

Double excitation of two atoms closer than $R_b$ is strongly suppressed — suppressed, not forbidden: the deeper inside $R_b$, the stronger the suppression (Challenge 01's second spacing sits just inside, where the blockade leaks). All three challenges run on that fact.

Device envelope

Baselines are stated against the analog device model in Pulser (pulser.AnalogDevice); the same limits are enforced on Pasqal Cloud. Verify Device.specs at runtime — the published envelope is authoritative, not this table.

Quantity	Symbol	Value	Notes
Interaction coefficient	$C_6/\hbar$	865,723 rad µs⁻¹ µm⁶	Rydberg level 60
Max Rabi frequency	$\Omega_{\max}$	12.57 rad/µs (2π × 2 MHz)	global channel
Max |detuning|	$\vert\delta\vert_{\max}$	125.7 rad/µs (2π × 20 MHz)
Max sequence duration	$T_{\max}$	6 000 ns	hard cap for all challenges
Min atom spacing	$r_{\min}$	5 µm
Max atoms / max radius	—	80 atoms / 38 µm from origin
Waveform clock period	—	4 ns	durations in multiples of 4 ns
Max runs per job	—	2 000 shots

Workflow & ranking

Workflow. Iterate in simulation; validate on Pasqal Cloud within your team's hardware-run budget (announced in Discord).

Ranking. Highest stage completed wins; ties within a stage are broken by that stage's own score.

The challenges

	System	Task	Score
Challenge 01	2 atoms	Bell-state prep under blockade, at two spacings	fidelity $F$
Challenge 02	4–5 atoms	embed a target graph, sweep to its maximum independent set	$P_{\mathrm{MIS}}$
Challenge 03	10–85+ atoms (hardware ≤ 80)	beat a published benchmark curve at matched instance size	paper's metric / $\mathcal{R}$

Submission format

Per team, in your Discord team channel by 16:30:

1. the highest challenge attempted, and for it: your waveforms and register (Pulser sequence, JSON, or code that generates them);
2. simulated score ($F$, $P_{\mathrm{MIS}}$, or $\mathcal{R}$) with the baseline value you beat, same shots;
3. Pasqal Cloud job IDs for your hardware validation runs;
4. three sentences: what you changed relative to the baseline, and why it worked.

Read ahead

• arXiv:1808.10816 — Quantum Optimization for Maximum Independent Set Using Rydberg Atom Arrays — the mapping; blockade radius as the unit-disk edge.
• arXiv:2403.11931 — Graph Algorithms with Neutral Atom Quantum Processors — graph problems in analog mode.
• arXiv:2511.22967 — Benchmarking neutral atom-based quantum processors at scale — the curve to beat.
• arXiv:2201.02773 — A Survey of Quantum Computing for Finance — background for the bonus.