Senior Staff Engineer at Google, building production ML systems at the intersection of large-scale learning and neuromorphic computing.
Borrow from biology only where it improves efficiency, adaptability, and signal fidelity in real systems.
Compute should react to events, not waste power waiting for them. The best systems wake up only when the signal matters.
A model is only valuable when it survives production traffic, hard latency budgets, and real failure modes at scale.
Representative systems across reinforcement learning, retrieval, model delivery, and neuromorphic runtime design. The through-line is production ML: high-scale inference, efficient adaptation, and architectures that survive real traffic.
A real-time decision engine for Google Ads that used reinforcement learning to optimize pacing, allocation, and market liquidity across multi-billion-dollar spend.
Retrieval and ranking systems for Google Drive and Ads that combined embeddings, nearest-neighbor search, and intent signals to surface the right result under tight latency budgets.
Production pipelines for training, deployment, and low-latency serving that kept models fresh in hours instead of days while maintaining reliability at global scale.
A low-power runtime for spiking neural networks that pushed event-driven inference closer to the sensor, reducing wasted compute and power on edge hardware.
Hover over data cards to decrypt technical definitions.
SPIKING NEURAL NET
REINFORCEMENT LEARNING
THE BOTTLENECK
LOCAL INFERENCE
Career Trajectory
Neuromorphic Computing & Hybrid AI
Applied ML & RL Infrastructure
Ranking, Retrieval & ML Platform
Computer Vision & Multimodal ML
Working on large-scale ML systems, neuromorphic computing, or production AI architecture? Let's talk.
Send Signal