Mathematics · Computational Science . Machine Learning · Education
Davian R. Chin FIMA
I study memory and scale in stochastic systems, build scientific machine learning tools, and bring mathematical ideas to life through teaching.
PhD researcher at the University of Reading
Scientific machine learning researcher & STEM educator
Reading, UK01 A LITTLE CONTEXT
My name is Davian Chin, a mathematical scientist, scientific machine learning researcher, and educator based in Reading, UK. My work brings together fractional calculus, stochastic dynamics, scientific machine learning, and computational neuroscience.
At the University of Reading, I investigate structural limitations of classical long-range dependence inference for heavy-tailed and non-stationary processes. I am developing a scale-organised, null-relative framework and reproducible benchmark suites, alongside GPU-accelerated generative models that make the assumptions, methods, and evidence easier to inspect.
Alongside research, I bring more than 15 years of teaching and leadership experience across the UK, China, Kazakhstan, and Jamaica. My teaching connects mathematical reasoning, computational thinking, and practical scientific enquiry, supported by curriculum design, teacher development, and Python and Jupyter learning resources. From the classroom to scientific code, I’m interested in making complex ideas both rigorous and accessible.
02 RESEARCH
How can we distinguish genuine dependence from apparent structure in complex data? My work connects multiscale mathematical foundations, null-relative inference, and reproducible scientific computation.
01 / INFERENCE
Long-range dependence in heavy-tailed and non-stationary processes. Developing a scale-organised, null-relative framework combining multiscale regularity, moment-free dependence measures, and aggregation-limit behaviour.
02 / METHODS
Fractional calculus and dynamical systems meet differentiable simulation, neural operators, and physics-informed learning. I build GPU-accelerated generative models to test methods under controlled departures from their assumptions.
03 / APPLICATIONS
Controlled fractional and scale-free generative models test whether dependence associated with neural criticality remains identifiable under heavy tails, non-stationarity, and observational distortion.
CURRENT DOCTORAL WORK
University of Reading · 2025–present · Expected completion 2028
My PhD investigates structural limitations of classical LRD inference and develops a scale-organised, null-relative dependence framework. I am formalising the Variance Trap, constructing moment-free and sub-quadratic dependence coordinates, and developing controlled generative models for neural applications.
03 SELECTED SOFTWARE
BENCHMARKING
A reproducible framework for evaluating long-range dependence estimators on synthetic, contaminated, and observational time series, with explicit assumptions and uncertainty reporting.
FRACTIONAL COMPUTING
Research-support software for fractional calculus and fractional dynamical systems, including JAX-native operators, Caputo solvers, and experimental differentiable workflows.
FRACTIONAL RESERVOIR COMPUTING
Fractional reservoir computing for critical brain dynamics, combining power-law memory, heavy-tailed drive, Besov regularisation, and a homeostatic threshold for quasi-self-organised criticality.
PHYSICS-INFORMED LEARNING
A focused research implementation of a Fourier neural operator for the 2D heat equation, exploring the balance between data-driven and physics-informed training objectives.
OPEN TEACHING RESOURCES
Mathematics lessons through Python and Jupyter notebooks, connecting functions, statistics, and mathematical reasoning with computational exploration.
CURRICULUM & LABS
A structured curriculum with learning outcomes, JAX and NumPyro labs, capstone projects, and assessment rubrics for scientific programming.
04 CURRICULUM VITAE
University of Reading · Reading, UK
Evaluated 15 classical, machine-learning, and neural LRD estimators. Developing null-relative dependence theory, moment-free mathematical methods, and differentiable fractional simulations for neural dynamics.
Neuryte Learning Ltd · United Kingdom
STEM education and AI consultancy, programme design, project delivery, and teaching across mathematics, science, computer science, and engineering.
Walton High · Milton Keynes, UK
Secondary computer science, computational thinking, and project-based learning for external examinations.
The Bedford Academy · Bedford, UK
Application-led teaching supported by simulation and interactive visualisation, with a focus on engagement and progress for lower-attaining learners.
Shenzhen Vanke Meisha Academy · China
Built and led a multicultural mathematics department, contributed to school strategy, and researched computational thinking in mathematics education.
Nazarbayev Intellectual Schools · Kazakhstan
Cambridge Integrated Mathematics, teacher development, and interdisciplinary summer STEM programmes in biodiversity and rocket science.
St Hugh’s High School · Jamaica
Curriculum and instruction for CSEC, CAPE, and A level courses; supervised mathematics projects and physics laboratory work.
Gaynstead High School · Jamaica
Mathematics instruction for students in Grades 10–11.
ACADEMIC BACKGROUND
2025–present · Expected 2028
University of Reading
Computational neuroscience and neural dynamics. Working thesis: Long-range dependence in heavy-tailed and non-stationary processes: theory, inference and neural applications. Focus on fractional function spaces, stochastic fractional neural dynamics, and physics-informed operator learning.
2023
University of East London
AI, machine learning, and big data technologies; physics-informed neural operators for heat and diffusion equations.
2018
University of South Wales
Development, management, and technology; computational thinking as a framework for mathematics education.
2008
University of the West Indies
Pure and applied mathematics, mathematical modelling, and operations research.
05 PUBLICATIONS & CONTRIBUTIONS
Journal article
D. Batic, D. Chin & M. Nowakowski · European Physical Journal C, 71, 1624
Talk & poster
School of Biological Sciences ECR Symposium · University of Reading · July 2026
Conference poster
School of Biological Sciences & Biomedical Engineering Research Conference · University of Reading
Manuscripts in preparation
A benchmark of classical LRD estimators on neural time series (target: NeuroImage / IEEE TBME); lrdbench: a reproducible benchmark framework for long-range dependence estimation (target: JOSS); The Variance Trap (target: Journal of Statistical Mechanics).
06 ARTICLES & TUTORIALS
I plan to publish regularly on research, STEM education, AI, and software engineering for scientific work, including occasional video tutorials.
COMING SOON
Expect research notes, practical STEM and AI education ideas, and explorations of scientific software. Published posts will appear here.
THE TOOLKIT
07 TUITION & RESEARCH RESIDENCY
Individual and small-group learning in advanced mathematics, computing, physics, and applied scientific modelling. Sessions are tailored to your starting point and the questions you want to tackle.
Calculus, linear algebra, differential equations, real and complex analysis, functional analysis, measure theory, and topology.
Programming, algorithms, software engineering, data science, machine learning, and reproducible scientific computing.
Physics and natural sciences, dynamical systems, control theory, and computational models in biology and neuroscience.
A short, guided mini-research programme in computational biology, physics, or scientific machine learning: formulate a question, build and test a model, and communicate the findings.
We can work through a course, build mathematical foundations for research, or design a focused programme around a specific problem.
Enquire about tuition08 STEM & AI EDUCATION CONSULTANCY
Through Neuryte Learning Ltd, I help educators and organisations design rigorous STEM learning and use AI thoughtfully in teaching and research.
Bespoke curricula, project-based learning, and enrichment programmes that connect mathematical ideas to computation and experimentation.
AI literacy, responsible classroom use, and learning experiences that put reasoning, evidence, and learner agency first.
Workshops, teaching resources, and support for embedding computational and AI tools into existing programmes.
Consultancy starts with your learners, aims, and constraints, then develops a practical programme or workflow you can use and evaluate.
Discuss consultancyRESEARCH · SOFTWARE · EDUCATION
For research collaborations, scientific machine learning opportunities, or conversations about STEM education, get in touch.