Quantitative Research & Systematic Modelling

I am a quantitative researcher with a background in statistical physics, stochastic modelling, and GPU-accelerated computation. I develop models and simulation tools to analyse and forecast complex systems.

WHAT I DO

Stochastic modelling & SDEs
I have extensive experience with stochastic differential equations, Langevin dynamics, and stochastic models with noise or disorder, all of which translate to price dynamics, volatility surfaces and risk processes.

Time-series & forecasting
I work with high-dimensional time-series, extracting structure and identifying regimes and transitions. In a finance context, I apply this mindset to signal extraction, feature engineering, and forecasting pipelines.

Simulation & scenario analysis
I design GPU-accelerated simulators that run orders of magnitude faster than naive implementations, allowing for large-scale Monte Carlo, stress testing, and scenario analysis.

Data-driven research workflow
From raw data to deployable code: I am used to building reproducible pipelines, cleaning and exploring data, choosing appropriate models, and validating results against reality.

SELECTED PROJECTS

Stochastic Interface Dynamics & Microstructure Modelling

Developed mathematical frameworks to analyse how complex interfaces evolve under noise, heterogeneity, and local interactions. Implemented methods for roughness quantification, memory effects, and dynamic response under perturbations.

Capabilities

Structure extraction in noisy systems; modelling path-dependent dynamics; identifying stability regimes.

Time-Series Modelling & Predictive Pipelines (ML + Statistical Methods)

Built end-to-end machine learning pipelines for forecasting and structural analysis of high-dimensional, noisy time-series. Work includes feature engineering, model selection, cross-validation, and stress-testing under realistic noise conditions.

Capabilities

Predictive modelling; regime detection; medium-horizon forecasting; scalable data pipelines.

High-Performance Monte Carlo for Stochastic Processes (CUDA/C++)

Designed and optimized GPU-accelerated solvers capable of simulating thousands of stochastic trajectories in parallel. Achieved orders-of-magnitude speed-ups over CPU-based methods through parallelization, memory optimization, and custom numerical integration schemes.

Capabilities

Large-scale scenario generation; uncertainty propagation; rare-event analysis; fast prototyping of stochastic models.

Uncertainty Quantification, Sensitivity Analysis & Model Robustness

Quantified statistical and systematic uncertainties in complex physical models. Analysed model sensitivity to noise sources, parameter variation, boundary conditions, and structural assumptions.

Capabilities

Robustness testing; model validation; stress-testing frameworks; assessment of model reliability under uncertainty.

PUBLICATIONS

TECHNICAL SKILLS

Modelling & maths

stochastic processes, SDEs, SPDEs, time-series, statistical learning, Bayesian inference, optimization.

Programming

Python, C++, CUDA, Linux, Git.

Data

Cleaning large datasets, feature engineering, exploratory analysis, visualization, and reproducible pipelines.

HPC

GPU acceleration (CUDA), parallelism, profiling, and optimization of numerical code.