Built by people who design
the math and ship
the machines.
Machine learning research meets the business that ships it.
Team Members
Cole Amaya
Cole designs and implements machine learning, deep learning, and LLM systems across research and applied product work. At Stokhos Labs, he leads model architecture, training pipelines, Neural SDE research, and engineering for the Fin-Tech risk analytics platform.
Email Cole →Filip Maleev
Filip studies Mathematics at Colorado College, where he competes in mathematics and works on distributed training systems for large language models. He leads Stokhos Labs' company strategy, customer development, and external partnerships.
Email Filip →Research-first.
Institution-grade.
Every architectural choice traces back to the underlying stochastic differential equation: drift and diffusion terms parameterized by neural networks, integrated with Euler–Maruyama, not a black-box sequence model retrofitted to look continuous.
The 64.3% result was never touched by in-sample optimization. We evaluate on data the model has never seen, under conditions it was never trained for. See the full backtest →
Built for any institution that carries interest rate risk (banks, insurers, asset managers, pension funds), not to impress an academic reviewer.
Work with us.
Any institution that manages interest rate risk: banks, insurers, asset managers, pension funds. Reach out to discuss a pilot, research collaboration, or demo.
Contact Both →Pilot Program
Access the full dashboard with your own data or our representative dataset. Run the Neural SDE alongside your current model for a direct MSE comparison.
Research Collaboration
Interested in applying Neural SDEs to other fixed-income instruments, credit spreads, or macro variables? We're open to joint research with serious institutional partners.