About
I'm Pierre Allain, an applied scientist based in London. Most of my work has been around NLP, language models, and large-scale text systems, and I'm especially interested in turning messy information into useful tools. I like building things that sit somewhere between research and engineering: technically serious, but made to be useful in the real world.
I grew up in France and studied in Paris, mainly at ENSAE, before starting my career in quantitative research in New York. I later moved to London and worked in a few different environments, including Checkout.com and G-Research. My path has taken a few turns, which in the end has been a good thing. It pushed me across different industries, ways of working, and technical problems, and helped me get clearer about the kind of work I value.
Over time, I found myself more and more drawn to language, text, and the challenge of making large amounts of information easier to explore and understand. That has led me into work on fraud models, retrieval systems, evaluation, and production NLP pipelines. What I enjoy most is taking something unclear or difficult to use and making it simpler, more reliable, and more useful.
Outside work, I spend a lot of time running and boxing, read quite a bit, and play guitar badly but happily. This site brings together some of the things I work on and some of the things I keep coming back to.