Joseph Michael Levermore
d Ingénieur En Spectroscopie Et Optique @Eden Tech
Signup · Get unlimited contacts
WORK HISTORY
d Ingénieur En Spectroscopie Et Optique @Eden Tech
Paris, FR
Led the development of SCOUT: an advanced particulate analyser identifying >4 µm particles via reflective microscopy, Köhler illumination, polarised light, and spectroscopic chemometric interrogation- Built custom computer vision models combining feature-text definitions with image-based training, achieving >90% detection accuracy; developed deep residual neural networks for spectral classification- Engineered micrometer-precision robotics and motorised stage control, integrating morphological and spectroscopic data into a unified analytical platform- Fine-tuned large language models to auto-generate regulatory-compliant reports and data visualisations for legislative submission- Authored patent-pending methodologies and managed all development pipelines via a self-built agile ticketing system.
EDUCATION
University of Plymouth
Bachelor of Science (BSc), Environmental Science
Lancaster University
Master of Science (M.Sc.), Biochemical and Environmental Toxicology
King's College London
Doctor of Philosophy (Ph.D) in Analytical Toxicology, Molecular Toxicology and Vibrational Spectroscopy
SKILLS
ABOUT JOSEPH MICHAEL LEVERMORE
I am a Senior Machine Learning Engineer and Physicist at Eden Tech. I specialise in architecting end-to-end AI systems: building the hardware, the software, and the models required to detect low-latency signals in high-noise environments. My background is in scientific analysis, but my daily work is engineering. I lead the development of SCOUT, a commercial high-throughput analytical platform. My role is to translate complex physical phenomena into robust code.What I\'m actually building:Multimodal Computer Vision: Created \'PlasticVision\' by fusing text embeddings with image data. Achieving >90% accuracy on unseen microscopic targets where standard CNNs failed.Signal Processing: Built Deep residual networks to isolate and predict specific chemical signals from highly noisy, stochastic spectral data.Robotics & Control: Wrote the control logic for micrometer-precision motorised stages, synchronising high-speed robotic movement with image data.Agentic AI: Deployed an autonomous AI agent that uses raw quantitative data to write full regulatory reports, reducing hours of manual drafting.I am driven by the challenge of \"Inverse Problems\"—using physics-informed machine learning to reconstruct ground truth from imperfect data. Whether in optical physics or predictive modeling, I focus on building systems that are mathematically rigorous and computationally efficient.I write about ML, Physics, and Code (Python and C++) here: https://pocoapoco.io
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.