Nicolas Legrand
Founder at lycid
Aarhus, Central Denmark Region
Summary
👤
Senior🎓
Top SchoolNicolas Legrand is a founder and senior researcher blending nine years of experience in computational cognitive science, probabilistic modelling, and agentic AI safety to build modular security layers for next-generation intelligent systems. Based in Aarhus, he bridges academic rigor—developing nonparametric Bayesian methods, novel interoception measures, and psychophysiological toolkits—with production-grade engineering in Python, Rust, and JavaScript. His open-source track record includes libraries for physiological signal analysis and contributions to core numerical tooling (notably improvements to the aesara math backend), reflecting a rare mix of theory and low-level implementation. He focuses on emergent failure modes of agentic systems such as belief drift and miscoordination, and on measurable, interpretable mitigations grounded in causal and biologically inspired frameworks. As a scientist-turned-founder, he applies insights from computational psychiatry and embodied computation to make agentic AI more robust and transparent in real-world deployments.
9 years of coding experience
5 years of employment as a software developer
Doctorat, Psychologie, Doctorat, Psychologie at Université de Caen NormandieLicence, Philosophie, Licence, Philosophie at Université Grenoble AlpesMaster, Sciences cognitives, Master, Sciences cognitives at Université Lumière Lyon 2
French, danois, English
Github Skills (12)
tensorrt10
automatic-differentiation10
tensor10
tensorflow10
python10
operation10
theano10
optimizers9
symbolic-computation9
optimizer9
compiler9
testing8
Programming languages (6)
Github contributions (5)
aesara-devs/aesara
Nov 2020 - Sep 2022
Aesara is a Python library for defining, optimizing, and efficiently evaluating mathematical expressions involving multi-dimensional arrays.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:24 reviews, 7 commits, 9 PRs in 1 year 10 months
Contributions summary:Nicolas contributed to the `aesara` library, a Python library for defining and optimizing mathematical expressions. Their work involved modifying core functionalities within the library, specifically focusing on the `Choose` and `Join` operations in theano/tensor/basic.py. They implemented improvements by integrating built-in Python functions and removing legacy components, improving code clarity and efficiency. The user also added test cases to validate the changes.
multi-dimensionalpythonmathematical-expressionsautomatic-differentiationoptimizing-compiler
embodied-computation-group/Cardioception
Oct 2019 - Feb 2023
Python Cardioception Package - Measuring interoceptive performance with Psychopy
Contributions:3 releases, 316 commits, 40 PRs in 3 years 3 months
metacognitionpythonmeasuringperformancepsychophysics
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.