About

I’m Yankl (pronounced YAHN-kul), or Jacob if you prefer. I’m a mathematician and researcher transitioning toward quantitative research, scientific computing, and AI safety/evaluation. I bring rigorous mathematical training to problems in numerical methods, algorithm design, and machine learning.

Currently I’m an AI Trainer at Handshake AI, authoring math problems and formal proofs for RLHF pipelines used to improve LLM mathematical reasoning. I also work on ML projects, including semi-supervised learning for historical document page segmentation and spectral analysis of dynamically evolving random networks.

Technical skills: Python (NumPy, SciPy, Pandas, Polars, PyTorch, scikit-learn), Julia (SciML, DynamicalSystems.jl), CUDA (beginner). I’m comfortable with numerical simulation, mathematical modeling, and implementing algorithms from research papers.

Research background: I completed my PhD at Stony Brook University under Misha Lyubich, specializing in holomorphic dynamics and hyperbolic geometry. My thesis studied one-parameter families of Schwarz reflection maps arising from Shabat-Belyi maps. After my PhD I was a Visiting Postdoc at TIFR Mumbai working with Sabyasachi Mukherjee, where I continued my research on Schwarz reflections.

Recent project: EvolvingNetwork is a Julia library for simulating stochastically evolving global networks and analyzing their connectivity and spectral properties.