RobinInce

Robin Ince


I work at the interface of research and product: statistical methods, machine learning, and production data pipelines for novel, high-dimensional signals.


My recent industry experience has been in early-stage deep-tech startups. Currently, I am a Staff Data Scientist at Apoha, where we characterise the properties of molecules and materials through their dynamic behaviour, using a novel biophysical sensing platform to accelerate R&D in drug discovery, food and consumer goods.


Previously, I was a Staff ML & Neural Scientist at Atlas, developing signal processing and machine learning for a consumer multi-modal brain sensing wearable, and a Senior Data Scientist at Fab AI, benchmarking large vision and language models for education.


Before this, I was a Senior Research Scientist (Research Associate Professor) in the School of Psychology and Neuroscience at the University of Glasgow, where I hold an honorary affiliation. My PhD is in Computational Neuroscience, and my research applied statistics, information theory and machine learning to complex data from electrophysiology, neuroimaging (EEG, MEG, fMRI) and behavioural experiments.


Gaussian-Copula Mutual Information (GCMI)

Efficient and robust estimation of mutual information from multivariate data

gaussogram

Fast adaptive time-frequency transform (Gaussian S-transform) with a Rust core and Python/MATLAB interfaces

Bayesian Prevalence

Estimating the prevalence of effects in a population from experimental data

AI for Education

Benchmarking vision and language models for education, and distilling LLM judgements into data-curation and reward models

parqview

Fast terminal UI for exploring Parquet files

Serverless Obsidian Sync + MCP

Self-hosted Obsidian LiveSync backend on Cloudflare's free tier (Workers, Durable Objects, R2), with an authenticated MCP server for AI agents

Partial Information Decomposition

Measures for decomposing multivariate information and entropy