Why TwoTail exists.
A letter from the founder.
I’ve always loved to measure things. My favourite toy as a child was my stopwatch. I’d throw paper aeroplanes out the window, and time how long they took to reach the ground.
Unsurprisingly, I ended up studying maths and statistics. There I discovered stochastic processes, Markov chains, metric spaces, and bayesian inference.
My career in tech continued the theme. I modelled the probability of winning a tennis match from any state. Piano recital metrics for pitch, rhythm, and tempo. And most recently price elasticity modelling for DTC e-commerce brands. Alongside building AI startups, I ran a small product analytics consultancy, helping software companies pick a Northstar KPI and the experiments to move it.
In 2025 my worlds converged. I brought my analytical mindset to the challenge of measuring AI, went deep into evals, and out the other side fell TwoTail.
Measuring agents well won’t be about tooling. It’ll be about technique. Who can figure out the right evals for a given agent application. Who can maximise token ROI, and tell correlation from cause. Who can assign credit for outcomes to actions.
These are the techniques I’m building TwoTail to make possible.
Come measure with me.
Taught data teams at.
I'm proud to have taught data scientists who have gone on to work for Europe's best tech companies.
Measure your agent properly.
Setup in 10 minutes. First calibrated judge within a week.