Sean Howell

researcher // engineer // technical leader

Los Angeles · San Francisco Bay Area

[ download resume (pdf) ]

about

Researcher and engineer focused on AI-driven scientific discovery. I build agents, models, and infrastructure that allow machines to participate meaningfully in physics and mathematics research and control complex systems, from research to production.

Recent work includes neural decoders and RL for fault-tolerant quantum computing, VLA models for autonomous experimental science, autonomous agent systems for physics research, formalization and verification, and compiler toolchains grounded in ZX-Calculus. Fifteen years in AI/ML, with R&D leadership across three quantum computing organizations.

B.S. Physics & Mathematics, University of Maryland, College Park.

experience

2024 — Present
Principal Engineer, AI for Quantum & Director, Quantum OS
PsiQuantum · Palo Alto, CA
  • Founded and leading AI for Quantum R&D: ML-based QEC decoders, RL for adaptivity in fault-tolerant quantum computing, scientific agents for physics research, multi-agent systems for calibration, control, and operation of quantum systems.
  • Building agentic discovery infrastructure for scientific AI: harnesses, trajectory capture, structured evaluation of AI-generated scientific artifacts.
  • Directed Quantum OS: designed scalable control plane, fabrics, OS software for utility-scale photonic QC. Grew engineering org by 15+ ICs YoY.
2025 — Present
Independent Research & Development — AI for Science & Systems
Los Angeles, CA
  • Developing architectures for AI-assisted scientific reasoning: evaluation, harnesses, provenance systems, and long-horizon agent frameworks for physics and mathematics.
  • Research at the intersection of reasoning models and formal methods, targeting formalization of results and open problems in quantum information science.
  • Structured representations of reasoning (typed graphs over claims, evidence, derivations, simulations, proofs) for verifiable multi-agent collaboration.
2022 — 2024
Software Engineering Manager
AWS Center for Quantum Computing · Pasadena, CA
  • Led engineering and research in quantum compilers, control software/hardware, QEC frameworks, and a distributed runtime for quantum information processing systems.
  • Built and mentored a team of scientists and engineers (0–30 years experience) developing scalable, fault-tolerant quantum computing systems.
2019 — 2022
Head of US Division
Q-CTRL · Los Angeles, CA
  • Led US R&D, software engineering, and product. Grew team 0→20 engineers, drove 10× YoY revenue growth, achieved 9000× algorithm performance improvement.
  • First demonstration of reinforcement learning for gateset design and calibration on a superconducting quantum processor. Pioneering work in AI for NISQ hardware.

projects

agents & research infrastructure
rustprivate
lem
A scientific agent for physicists that unifies an Aletheia-style GVR loop and RLM techniques.
pythonpublic
deep-gvr
A minimal generate-validate-revise loop for Hermes-Agent, inspired by DeepMind’s Aletheia agent for autonomous research with empirical, formal, and human verification.
rustprivate
skaffen-mono
A general-purpose personal research, development, engineering, and inquiry agent.
rustprivate
culture
Multi-agent orchestration and AI research assistants.
pythonprivate
parallax-science
A platform and protocols for AI agents to engage in structured scientific discourse.
texprivate
hilbert
Solving open problems in quantum information science using AI.
zx-calculus & fault-tolerant quantum computing
rustprivate
loom-zx
A ZX-Calculus focused proof-carrying FTQC platform.
pythonprivate
zx-webapp
An interactive webapp for exploring fault tolerance for quantum computing in the ZX-Calculus framework.
quantum error correction & decoders
pythonpublic
nano-qec
Autoresearch-inspired neural decoder training using Hermes Agent and Stim.
pythonprivate
sabaki
RL training for QEC pre-decoders using Pufferlib, Stim, PyMatching, and sinter.
pythonprivate
open-qvla
Open vision-language-action models for quantum control, calibration, and error correction.
systems & platforms
rustprivate
bandalong
Authority plane reference implementation in Rust.
goprivate
chapterhouse
A reference intelligence plane implementation.
pythonprivate
pelagos
A modern, scientist-focused platform for oceanography.

writing

2026-04
Doing ML to Build Quantum Computers to do ML With →
On the recursive structure of machine learning and quantum computing, and why a new result on quantum oracle sketching gives the second half of the loop a rigorous foothold.
2026-03
On the Microfauna of Megaprojects →
Why the biggest machines of the next decade will be inhabited by dense ecologies of small machine intelligences, and why the organizations that matter should start cultivating them now.
2026-02
AI as the Control Plane for Impossible Machines →
Why utility-scale quantum computers, autonomous laboratories, advanced fabs, and large robotic systems will need advanced AI built into the machine as an operating layer.
2025-05
We’re growing up (face it, you’re young) →
A short note on the spirit of the era, written to no particular age of reader, chronological, spiritual, biological, or cybernetic.

publications

  1. H. Levine, A. Haim, et al. Demonstrating a long-coherence dual-rail erasure qubit using tunable transmons. Physical Review X, 2024.
  2. Y. Baum, M. Amico, S. Howell, et al. Experimental Deep Reinforcement Learning for Error-Robust Gateset Design on a Superconducting Quantum Computer. PRX Quantum, 2021.
  3. Y. Baum, S. Howell, et al. Reinforcement Learning for Error Robust Control on Cloud-Based Superconducting Hardware. Bulletin of the American Physical Society, 2021.

contact