Software architect and technical lead. I lead the technical side of a low-code generative AI product at Accenture: shaping the architecture and design with the team, and reviewing the core of what the wider project ships. I also work on the governance of agentic systems, evaluating them against the NIST AI RMF and the EU AI Act. Before that I helped build a quantum computing practice from nothing. Further back, a doctorate in theoretical physics, which is still how I think.
Most of what I build is client work and lives in private repositories. What's public:
Accenture/Q-Opt-suite — a benchmarking suite for quantum and quantum-inspired optimisation solvers. Compares D-Wave, Toshiba SQBM+, Quantagonia and AWS Braket platforms against a Gurobi classical baseline, across bin packing, max-cut, TSP, feature selection and spin-glass problem sets. Python, MIT. This is the framework behind my Conf42 talk on the state of quantum annealing.
Writing — Tensor Neural Networks Significantly Cut Computational Cost of Low Latency Object Detection in Satellite Imagery, lead author, on applied research at Accenture's Centre for Advanced AI. Earlier: two co-authored Java titles for Wrox and Apress, and a doctoral thesis on QED renormalization in anti-de Sitter space.
More, including how to reach me, at peterdenhaan.com · LinkedIn