Hi, I'm Suyeol Yun, Head of AI at LinqAlpha.

My job is to equip the LinqAlpha team with the right framework for understanding where AI is heading, and to figure out where we should stand to be ready for it. In practice, the work comes down to one thing: closing the loop. I build the evaluation pipelines that tell us whether we're genuinely improving, I put coding agents to work climbing against those signals, and I pull our teammates into the same loop, so that we can improve the system with confidence.

I studied mathematics as an undergraduate and began my career as a deep learning engineer. For my first four to five years I worked in computer vision, mostly on facial reenactment: triangulating hundreds of photographs shot on a lighting stage into 3D meshes, and designing a graph neural network (GNN) to transfer facial expressions from one person's sequence of meshes onto another's. It was a hybrid of two worlds. On one side, the deterministic world of graphics, all mountains of C++ and the Eigen library, where everything has to line up exactly. On the other, the empirical side of deep learning, where so much comes down to testing: tuning dropout, deciding where to place residual connections, and so on.

That GNN work eventually led me to the U.S. lobbying industry, and from there into a PhD in political science at MIT. But I dropped out in 2022, when ChatGPT arrived. I joined LinqAlpha because it was the right place to study, empirically, how institutional investing actually works.

I keep trying to define my role at LinqAlpha as that of a philosopher. Think of Jean-Jacques Rousseau, arguably the most significant intellectual influence on the French Revolution, who supplied the ideological blueprint for its democratic ideals. He challenged traditional authority and proposed instead a society driven by the "general will" of the people. In that same spirit, I try to give my teammates the most intuitive framework for making sense of this surge of AI agents we're living through, and the strategic thinking it demands.

I love to debate and to be challenged, and I love watching our culture take shape through those debates. We roll the dice to make the right move, constantly optimizing our estimate of the value function. As Thompson sampling reminds us, the best action is never a fixed one; we should shape our policy to match the uncertainty of the world. So my job isn't to exploit what we already know. It's to quantify, and qualitatively articulate, what we don't. And then to build a stochastic policy as a team that wins the game we're playing.