WildtConsulting
Engineer, researcher, founder — from race cars to Edge AI in the wild, turning state-of-the-art science into products customers love.
Background
My research sits at the intersection of machine learning and simulation science — Bayesian inference, mathematical modeling, and stochastic methods, grounded in scientific computing and applied statistics. In practice that means work like Gaussian processes and arbitrary polynomial chaos expansion, Bayesian model calibration, surrogate models, contaminant transport, and optimal control.
Day-to-day I build with Julia and Python. I'm at home on Google Cloud and AWS, and I've shipped applied ML — keyword spotting and sound event detection — on edge devices. And on the people side: I ask, and I listen.
The race cars? Before any of this, I was on the aerodynamics team of University of Waterloo Formula Motorsports — where I learned that simulations are only worth as much as the lap times they predict.
What I do
Bayesian modeling & uncertainty quantification (UQ)
Gaussian processes, polynomial chaos, model calibration, surrogate models — answers with error bars.
Edge audio ML
Keyword spotting and sound event detection that run on-device — small, fast, private.
Scientific computing
Research-grade numerics turned into production code, in Julia and Python.
Cloud & product
From prototype to deployed product on Google Cloud and AWS.
Selected work
Companies
auditect — a startup I co-founded tackling the care crisis with AI-powered ambient monitoring. Our system detects falls and emergencies in real time, cutting response times and freeing caregivers to focus where it matters most.
Web applications
SocraticFinance — retirement-planning tool helping people understand and manage their own investments. Built together with the SocraticFinance research team.
timriddle — stochastic simulation of the German TV show Gefragt — Gejagt.The German equivalent of "Beat the Chasers" — solo contestants against an expert chaser, played out probabilistically.
Open-source libraries
FastARD.jl — fast automatic relevance determination in Julia.
SteinVariationalGradientDescent.jl — SVGD inference in Julia.
SimpleKernelregression.jl — kernel regression in Julia.
Elsewhere
Research profile — University of Stuttgart, Dept. of Stochastic Simulation and Safety Research for Hydrosystems
→ Reach out if you'd like to collaborate, or think I could add value to your product.