Anthony Polloreno, Ph.D.

I'm a mathematically-oriented engineer with over a decade of experience in software engineering, machine learning, and quantum computing. My work typically explores the structure and characterization of information-processing systems.

I'm currently at Apple, building large-scale machine learning infrastructure. Prior to that, I was at Essential AI, where I trained, optimized, and deployed large language models (we co-developed Fin)!

Previously, I was a postdoc at JILA and CU Boulder, focusing on how noise affects learning dynamics—particularly in recurrent networks. Before that, I worked with Graeme Smith at JILA and the QPL at Sandia National Laboratories during my Ph.D. at CU Boulder, researching machine learning and quantum information.

Earlier in my career, I was a software engineer at Rigetti Computing, developing quantum control systems and working with Sandia on benchmarking quantum computers. I have a triple major in computer science, mathematics, and physics from UC Berkeley, and did research at both LBNL (reflection zone plates) and Sandia (quantum control theory).

---

I also write papers:

• Practical Efficiency of Muon for Pretraining
• Rethinking Reflection in Pre-Training
• Limits to Analog Reservoir Learning
• A Theory of Direct Randomized Benchmarking
• A Note on Noisy Reservoir Computation
• The QAOA with Few Measurements
• Opportunities and Limitations in Broadband Sensing
• Individual Qubit Addressing of Rotating Ion Crystals
• Towards Demonstrating Fault Tolerance in Small Circuits Using Bacon-Shor Codes
• Robustly Decorrelating Errors with Mixed Gates
• Parametrically-Activated Gate Protected from Flux Noise
• Quantum Kitchen Sinks
• Universal Parametric Entangling Gates
• Parametrically Activated Entangling Gates with Transmons