Multi-GPU Quantum Simulation
Distributed statevector simulation with NVIDIA CUDA-Q across multiple GPUs and nodes on Vast.ai.
Qiskit Advocate · IBM-recommended
Quantum & AI research that turns into working systems.
I help teams ship quantum algorithms, GPU-scale simulation, and applied AI — from first feasibility study to reproducible, production-ready delivery.
- Multi-node multi-GPU CUDA-Q simulation
- 15+ verified quantum credentials · 5 accepted UnitaryHack PRs
- Building LLMs since 2020 — before ChatGPT
Trusted across IBM, the Qiskit community, and open-source quantum software.
Track recordVerified
Quantum credentials (Credly)0
Accepted PRs · UnitaryHack 20260
Qualification Round 2026#1 prelim
Languages in 2020 GPT model0
Qiskit Advocate & QGSS mentorYes
Qiskit Advocate
IBM-recommended
15+ Credly badges
CUDA-Q · Multi-GPU
UnitaryHack 2026
Services
Three ways to work together
Clear scopes, fixed milestones, reproducible results. Every engagement ends with documentation and benchmarks.
Quantum R&D
Algorithm design, circuits, and feasibility studies for real quantum hardware and simulators.
- Qiskit / circuit design
- Feasibility & literature review
- Prototype + benchmark report
GPU / HPC Simulation
Scale quantum simulation across multiple GPUs and nodes with NVIDIA CUDA-Q.
- Multi-node statevector sim
- Cloud GPU orchestration
- Performance tuning & scaling
Applied AI / LLMs
Practical machine-learning and language-model builds — fine-tuning, evaluation, deployment.
- LLM fine-tuning (PyTorch)
- Multilingual NLP
- Evaluation & deployment
Selected work
Proof, not promises
Recent projects across quantum HPC, language models, and open-source — each independently verifiable.
gpt2-small-czech-cs
Fine-tuned GPT-2 language model for Czech, open-sourced on GitHub — reproducible NLP.
Multilingual GPT — 2020
A ChatGPT predecessor on GPT-2 with CS/EN/AR/PL/DE translation — years ahead of the curve.
UnitaryHack 2026 — 5 PRs
Five accepted pull requests to open-source quantum software during UnitaryHack 2026.
Qualification 2026 — Winner
Top of the preliminary scoreboard in the 2026 quantum qualification round.
Qiskit Advocate & QGSS
Qiskit Advocate; completed all QGSS dev-week labs; mentoring at QGSS 2026 this July.
How it works
A clear, low-risk path
Four steps from first call to delivery — predictable scope, predictable cost.
1
Discovery
Scope the problem, feasibility, constraints, and success metrics.
2
Design
Algorithm/model design, prototype, and a milestone plan with a quote.
3
Build
Reproducible, tested code with real backend or multi-GPU integration.
4
Deliver
Docs, benchmarks, handover — plus optional ongoing support.
0
Quantum credentials
0
Accepted PRs · 2026
#1
Qualification (prelim)
2020
Building LLMs since
What people say
Backed by industry & community
★★★★★
"George pairs genuine quantum depth with the engineering discipline to actually ship. A rare combination."
★★★★★
"His CUDA-Q simulation scaled cleanly across nodes — exactly the HPC-quantum engineering the field needs."
★★★★★
"Five accepted PRs in one hackathon. He understands open-source quantum software and how to contribute."
FAQ
Common questions
Quantum algorithm R&D, GPU/HPC quantum simulation (CUDA-Q), and applied AI/LLM builds. Feasibility studies, prototypes, and production-ready delivery — plus QGSS 2026 mentoring.
Yes — as a Qiskit Advocate I design for real backends and high-fidelity simulators alike, and scale statevector simulation across multiple GPUs and nodes when hardware limits bite.
A short discovery call to scope the problem, then a written plan with milestones and a fixed quote. You approve before any build work begins.
Reproducible, tested code, clear documentation, and a benchmark report — with optional ongoing research support.
Have a hard quantum or AI problem?
Tell me what you're trying to build. You'll get a clear, honest assessment — and a plan to get there.
Get a quote
Most replies within one business day. Prefer a channel? Reach me directly below.