Fatih Kaya.

My research journey began with a deceptively simple question: How do you teach an algorithm a truth that even humans sometimes fail to perceive? Initially, this was more of a curiosity than research for me. At that time (2018), I was in my first year of undergraduate studies and had begun to explore newly emerging image processing techniques. A photograph was being converted into object labels in seconds... at the time, it seemed like alien science to me. I researched these topics for a long time, but I still couldn't convince myself that it wasn't alien science. I didn't publish my work on this as a paper, frankly, my aim at the time was simply to explore:

There's a quiet, delightful peculiarity to bestowing an imaginary life upon steel. The digital twin is less a schematic and more a shadow bound to bone and iron, born and aging alongside its twin so that the factory grounds don't collapse and disintegrate under our watchful eye. It captures a rhythm where human vision often falls short, saving us from the sudden, ear-piercing groans of weary machines and the blind panic of unexpected stops. But beyond merely anticipating wear and tear, it imbues the cold machinery with a kind of borrowed instinct. In those rare moments of autonomy when we take our hands completely off the wheel and trust the system to find its way within its own self-defined world, this shadow becomes an internal compass. If the machine can survive here, it's only because its twin has already traversed each of those dark and difficult corners long before the first gear even begins to turn.

Each year the city grows more crowded, millions of people simply want to get to work, get home, or sit by an open window before nightfall. But our desire to move has long surpassed the desire to pave the world with asphalt. When physical space runs out, the solution ceases to be a matter of asphalt and becomes an architecture of thought. This is where computer science comes in, not with bulldozers, but with logic. It treats the expanding urban labyrinth as a giant, living graph, allowing algorithms to sense the weight of our restless haste long before it turns into congestion. Through real-time optimization, dynamic routing, and distributed coordination, lines of code silently absorb the friction of an overcrowded world. It anticipates fluctuations, balances competing currents along quiet paths, and directs an overwhelming wave of people into an effortless flow. This is a subtle, invisible blessing.

A solitary mind, however brilliant, always runs up against the ceiling of its own breath. For years we tried to build giants, monolithic intelligences tasked with carrying the entire weight of our ambition alone, only to find that a single center of gravity carries too much inertia to outrun the horizon. Multi-agent systems abandon the romance of the solitary thinker. Instead of one lonely oracle, they assemble something like a multitude of agile, specialized wills, each holding only a fragment of the world, debating and negotiating across silent channels in a dialect of pure intent.

When we accelerate that collective pulse, stripping the friction from their consensus and shaving the hesitation from their shared memory, thought ceases to be a slow, linear march. They divide the impossible into microscopic duties, fanning out across thousands of hypothetical futures in the quiet space between heartbeats. It is not just about computing faster, it is also about folding time itself, letting a choir of artificial minds solve tomorrow’s collisions before we have even finished pointing the way.

And right there, that is where my favorite research comes in. Trying to understand humanity, which may very well be impossible, but perhaps that was never the failure. Perhaps that was always the point...

We spent generations asking computer science to tame our nature, feeding our unruly instincts into crisp conditional statements and expecting a human soul to compile cleanly. But we are not invariant states. We are creatures who will choose cobalt over crimson on a sudden, inexplicable urge, order a bitter drink on a rainy afternoon just to feel the sting, or rewrite our entire lives because the autumn light hit an alleyway at an unfamiliar angle. You cannot cage that in a deterministic if-else branch without suffocating the very thing you were hoping to observe.

In the end, this research isn't about writing code that solves humanity like an equation. It is about teaching silicon how to leave room for the mystery.