Alek Hunter Kemeny

Dual-Species Rydberg Atom Array for Quantum Computing and Simulation

Harvard University Physics | PhD Student, Advisor: Professor Kang-Kuen Ni | September 2024 - Present

Building a neutral atom quantum computer with cesium and sodium atoms. I completed the design and construction of a new ultra-high-vacuum chamber. This upgrade positions the experiment to reach higher Rabi frequencies on sodium and suppress stray-field-induced dephasing, key parameters for the high-fidelity quantum operations we plan to demonstrate in the coming year. In parallel, I wrote FPGA-based control code that automatically locks our excitation lasers, eliminating daily manual tuning; I also installed a new optical breadboard for the improved imaging fidelity of sodium. Ultimately, I aim to use AI and automation techniques to build a self-driving lab.

Inkwell, Inc. The First AI-Native Classroom

Cambridge, MA | CEO and Co-Founder | June 2023 - Present

I co-founded Inkwell with the aim of providing ultra-personalized education for students and supercharging teachers with AI grading & analytics. I first co-led a team of 10 open-source developers building generative AI tutors that generate Courses, Labs, and MIT course content. I've organized Laboratory in Quantum Systems Engineering and BT05 Photonics Bootcamp. I realized that bolting chatbots onto existing courseware left powerful capabilities on the table, so I have invested the past several months in building Inkwell, the next generation of EdTech. I currently have $300k revenue and are working with our clients at the NSF and Center for Semiconductor Manufacturing to deploy Inkwell in classrooms this fall. I was invited to present this work at MIT Generative AI Week¹ as well as the Fairness with ChatGPT Workshop² at the University of Chicago.

IBM Quantum Compilers and Benchmarks

MIT-IBM Watson AI Laboratory | Full-Time Research Engineer, Manager: Dr. Kevin Krsulich | September 2023 - September 2024

Built a novel benchmarking tool for quantum compilers. As a Qiskit Developer, I also fix bugs, add features, and manage Qiskit releases. As an intern in summer 2022, I created a software package, spinlocker, that serves as an extension for Qiskit experiments. This package contains Nuclear Magnetic Resonance-inspired pulse gate experiments (T₁ρ) that I programmed for benchmarking the coherence times of single and entangled qubits on IBM superconducting quantum hardware (Python, Qiskit, IBMQ).

Utilizing Entangled Quantum Sensor Networks for Navigation Without GPS

MIT RLE | Researcher, PI: Professor Dirk Englund | June 2023 - Present

Developed a procedure to enable the navigation of vehicles in GPS-denied locations. I developed a general representation for dynamic, distributed magnetic fields and proved a Quantum Cramer-Rao bound for magnetic field estimation. I created an algorithm for optimal placement of entangled quantum sensor networks. I currently supervise an undergrad who is implementing an algorithm to use my optimal measurements for precisely determining the location of our vehicle anywhere on the planet using a map of Earth's magnetic field.

ML Pulse Optimization and Simulation of Trapped-Ion Quantum Computers

Duke & Sandia National Labs | Researcher, PI: Professor Kenneth Brown | June 2021 - August 2023

Built the first AMFM pulse optimizer and trapped-ion simulator. I implemented a machine-learning pipeline to find robust AMFM pulse sequences for user-defined gates. The project tackled optimization with Sandia National Laboratories, where I was the first to integrate Sandia's trapped-ion hardware. I also implemented a novel ML method achieving state-of-the-art fidelity (99.92%) and robustness for the Molmer-Sorensen gate by solving the master equation at every step in optimization to incorporate a complete picture of system noise.

Laser-Locking for Improved and Consistent Gate Times in a Trapped-Ion Quantum Computer

Duke Quantum Center | Researcher, PI: Professor Kenneth Brown | December 2022 - May 2023

Built hardware to achieve consistent gate times in a trapped-ion quantum computer. More specifically, I constructed a circuit that performs a PID loop for locking the amplitude of our pulse gates. I aligned the laser beam by manually tuning mirrors, soldered previously broken voltage controllers, and break-point tested using a spectrum analyzer. Ultimately, I was successful in locking the laser, confirmed by calculating the Allan deviation.

A Superconducting Circuit for Generating and Studying Slow Light

Princeton University | QuNIP Intern, PI: Professor Andrew Houck | June 2022 - July 2022

Analytically and computationally investigated superconducting circuits that produce slow light phenomena, called "non-Markovian" physics. I developed a rigorous mathematical argument constructing the minimum area circuit that can generate slow light behavior (Mathematica). I implemented a time-evolution calculation of the circuit coupled-oscillator model and started predictable double peak behaviors from simulations (Python, NumPy, Matplotlib). Experimentalists plan to build this circuit on superconducting quantum chips to provide a new quantum simulator for slow light.

AI Image Recognition Algorithm for State-of-the-Art Particle Tracking @ Large Hadron Collider

CERN | Researcher, PI: Professor Ashutosh Kotwal | June 2021 - June 2023

Implemented novel AI algorithm for real-time particle tracking and identification. I optimized our algorithm for FPGA integrated circuit architecture (C++, Xilinx Vivado) to achieve a throughput of ~25 ns, allowing for fragment path reconstruction at 40 MHz. This constitutes a three order of magnitude increase in speed over current LHC architecture for data beats, state-of-the-art performance for image recognition that may correspond to dark matter. The ultimate goal of the project is to upgrade computer systems at the LHC to search for the lightest supersymmetric particle tracking!