Bio

Hi, I'm Ayesh. I build neuromorphic and brain-inspired AI systems designed to run under strict power, latency, and reliability constraints. My work spans spiking models, hardware-aware learning, and end-to-end software stacks—including compilers and runtimes—that translate research into deployable systems.

I'm a founder shaped by production systems, and I believe AI progress will come from efficiency, robustness, and specialized edge models—not ever-larger models.

My Work & Research

My work focuses on building AI systems that remain effective under hard constraints on compute, power, and reliability. I take a neuromorphic, brain-inspired approach and treat hardware as a first-class design variable rather than a deployment afterthought. This shifts the abstraction stack: from spiking model representations to event-driven intermediate representations and execution layers that interface directly with specialized chips. At this level, choices around time, sparsity, and scheduling are not implementation details—they directly determine what learning and inference regimes are even possible when latency and energy budgets are fixed.

This perspective is grounded in production reality. Before my current work, I built and deployed edge ML and computer vision systems for industrial customers at a YC-backed company, where robustness and uptime mattered more than leaderboard metrics. As a founder and CTO, I continue to operate at the boundary between research and deployment, using real systems to pressure-test ideas. The goal is not theoretical elegance, but architectures and software stacks that actually survive in the field—where efficiency, reliability, and long-term deployability are non-negotiable.

My Background

Outside of work, I write regularly—blogs and social posts on X and Instagram—sharing thoughts on the industry, art, poetry, and life. I was a founding member, donor, and volunteer with NUST Animal Rescue. I also have a deep appreciation for music and the arts; they keep me grounded and remind me there is more to the human experience than the systems we build.