About Me
Welcome to my personal website!
I'm a Machine Learning Robotics Engineer with 4+ years of experience in reinforcement learning, autonomous systems, and neuromorphic computing. I'm passionate about building intelligent agents that interact with the physical world, with a track record of deploying models to embedded systems and real-time hardware.
Currently, I'm an AI Engineer and researcher at Harbor AI in New York City, where I develop models for dynamic insurance price optimization and build scalable MLOps infrastructure. I recently completed my Master's in Aerospace Engineering at TU Delft and conducted visiting research at Harvard University's Edge Computing Lab.
I will be presenting my work on adaptive surrogate gradients for sequential reinforcement learning in spiking neural networks at NeurIPS 2025, if you will be attending, please feel free to reach out!
Projects
A2C using Spiking Actors and Critics
Implementation of Actor-Critic reinforcement learning algorithms using spiking neural networks, exploring the advantages of neuromorphic computation for energy-efficient RL agents.
Control using Spiking Neural Networks for Drone Landing
Real-world application of spiking neural networks for autonomous drone landing control, building on A2C research and presented at ICNCE 2024.
Neuromorphic Benchmarking using NeuroBench
A closed-loop tutorial demonstrating neuromorphic benchmarking techniques using the NeuroBench framework for fair and representative evaluation of neuromorphic algorithms.
A2Perf: Real-World Autonomous Agents Benchmark
Comprehensive benchmarking suite for autonomous agents with real-world domains including chip floorplanning, web navigation, and quadruped locomotion. Accepted at TMLR.
Adaptive Surrogate Gradients for SNNs - Blog Series
A 3-part blog series on training spiking neural networks with adaptive surrogate gradients. Covers gradient effects, sequential RL algorithms, and real-world deployment on the Crazyflie drone. Accepted at NeurIPS 2025.
Curriculum Vitae
Experience
Lead Machine Learning Researcher and Engineer
Harbor AI
Aug. 2024 – Present
New York City, USA
- Leading the development of reinforcement learning models for dynamic insurance price optimization
- Conducting interpretability research for inherently explainable risk analysis systems
- Building and scaling the MLOps infrastructure to support company growth beyond $100M valuation
Visiting Researcher, Machine Learning
Edge Computing Lab, Harvard University
Aug. 2023 – Jul. 2024
Boston, USA
- Co-authored and maintain NeuroBench, a benchmarking framework for neuromorphic and non-neuromorphic algorithms; contributed metrics, publication writing, and community engagement
- Contributed to A2Perf, a reinforcement learning benchmarking suite; ensured cross-platform compatibility and reproducibility
- Explored RL-based pruning methods for resource-efficient drone controllers using TinyML and event-driven neuromorphic computation
Machine Learning Researcher
Honours Programme MSc, TU Delft
Aug. 2020 – Dec. 2024
Delft, NL
- Developed and deployed neuromorphic MAV controllers using event data and optic flow, trained via reinforcement learning with on-device fine-tuning
- Designed implementations of deep Q-learning (DQN) algorithms for low-power RL on edge devices
- Created structural topology optimization tools for compliant wing morphologies in aerospace applications
Structural and Control Systems Engineer
AeroDelft (Hydrogen Aircraft Project)
Sep. 2022 – Dec. 2023
Delft, NL
- Led design and implementation of control systems in C++ for the world's first student-built hydrogen-powered aircraft
- Prototyped and translated CAD designs into fully tested flight-ready hardware
Education
M.Sc. (+ Honours Programme) Aerospace Engineering
University of Technology Delft
Sept. 2022 - December 2024
Delft, NL
Specialization: Control and Simulations
B.Sc. (+ Honours Programme) Aerospace Engineering
University of Technology Delft
Sept. 2019 - Aug. 2022
Delft, NL
Publications
NeuroBench: Advancing Neuromorphic Computing through Collaborative, Fair and Representative Benchmarking
Nature Communications, NICE Conference 2024
NeuroBench: Closed Loop Benchmarking
Neuromorphics Netherlands 2024, ASPLOS 2025
A2Perf: A Benchmarking Suite for Evaluating Autonomous Agents in Real-World Domains
Under review, AAMAS 2026
Control with Spiking Neural Networks Trained with Reinforcement Learning Using Surrogate Gradients
ICNCE 2024
Adaptive Surrogate Gradients for Sequential Reinforcement Learning in Spiking Neural Networks
Oral Presentation at NeurIPS 2025