
Liam Moore
Verified Expert in Engineering
Data Scientist and Developer
Leuven, Belgium
Toptal member since July 11, 2022
Liam is a machine learning scientist and engineer from a particle physics research background with experience leading various industry projects. He specializes in building bespoke statistical models, either using modern deep learning algorithms or classical machine learning combined with domain knowledge. Liam focuses on applying, developing, and keeping up to date with modern research while working with clients who care about creating social value.
Portfolio
Experience
- Python 3 - 10 years
- Machine Learning - 6 years
- Computer Vision - 6 years
- Deep Learning - 6 years
- GIS - 3 years
- Generative Pre-trained Transformer 4 (GPT-4) - 1 year
- LlamaIndex - 1 year
- LangChain - 1 year
Availability
Preferred Environment
Linux, Python 3
The most amazing...
...thing I've improved is the selection of viable embryos in fertility clinics using AI.
Work Experience
Senior AI Scientist
SONY
- Developed web security penetration tools using large language models.
- Analyzed academic research trends across adversarial machine learning and kept stakeholders abreast of essential developments affecting key areas of their activity.
- Investigated applications of AI across relevant market segments and co-designed mitigation protocols for adversarial threats.
Lead Data Scientist
CARE Fertility
- Designed and developed a bespoke deep computer vision and time series analysis solution for producing expert human-level time-lapse video annotations in the context of embryo incubation.
- Supervised the junior and mid-level data scientists and assisted engineers in productionizing and deploying the models in a user-facing cloud-native application.
- Beat the previous state-of-the-art product on the market for solving the same task.
- Co-published an abstract for work presented at the conference of the European Society of Human Reproduction and Embryology.
Senior Data Scientist
Unilever
- Developed a real-time anomaly detection model running on an Azure Stack Edge device in a factory.
- Trialled a combination of supervised/unsupervised modeling approaches using anomaly detection, object detection, and transfer learning.
- Worked with the staff on the ground on camera positioning, data acquisition, and efficient labeling strategies to maximize performance.
Senior Data Scientist
GIM
- Applied computer vision and time series analysis to remote sensing, earth observation raster, and vector data.
- Built and deployed a cutting-edge deep semantic segmentation model that identified buildings and objects from public orthophotographs across the Benelux region.
- Published a report for the cartographical administration of Luxembourg, detailing the algorithms and pioneering methods of quantifying the quality of the results when vectorized and used for constructing a database of buildings.
Postdoctoral Researcher
Université Catholique de Louvain
- Developed and compared computer vision and physics-aware deep learning models applied to radiation jet tagging.
- Performed statistical analyses of novel physical phenomena affecting the spin of top quarks to estimate the values of unknown physical parameters.
- Developed a tool for automatically simplifying effective quantum field theories in Mathematica.
- Supervised the PhD students on several research projects.
Marie Curie Early-stage Research Fellow
Cern
- Developed a Mathematica application that simplifies effective quantum field theories by removing operator redundancies.
- Performed statistical sensitivity analyses to novel spin correlations in measurements of top-quark pair production at the LHC.
- Collaborated with the experimental physicists, organizing the self-annihilation working group.
PhD Researcher
University of Glasgow
- Performed the research on constraining the parameters of the standard model effective field theory (SMEFT).
- Co-founded the TopFitter collaboration, publishing a series of papers performing the first global fit of SMEFT parameters to top-quark measurements at the LHC and Tevatron.
- Developed the Mathematica model-building tools used by the community.
Experience
AI-assisted Embryo Selection
https://www.eshre.eu/ESHRE2022/Programme/Searchable#!abstractdetails/0000694000I was the lead data scientist from the initial literature review through modeling on AzureML to productionization and deployment on AWS. I led a small team designing a solution that used multiple models with auxiliary objectives and custom loss functions accounting for subjectivity to tame a dataset of 450 million images.
Enriching Digital Maps with AI
https://business.esa.int/projects/eo4belmapWe determined the approximate years of construction of all buildings in Belgium using deep computer vision and time series analysis on public orthophoto and satellite data. We also obtained high-quality building footprints, which populated a database used in user-facing map products.
Datamining Objects from Public Earth Observation Imagery
https://data.public.lu/en/datasets/extopia-extraction-dojects-topographiques-par-intelligence-artificielle/I was the lead data scientist from the initial literature review through modeling on Azure ML to productionization and deployment on AWS. I led a small team designing a solution that used multiple models with auxiliary objectives and custom loss functions accounting for subjectivity to tame a dataset of 450 million images.
Production Line Fault Detection with Computer Vision
Education
PhD in Physics
University of Glasgow - Glasgow, Scotland, UK
Master's Degree in Physics
University of Glasgow - Glasgow, Scotland, UK
Bachelor's Degree in Physics
University of Glasgow - Glasgow, Scotland, UK
Skills
Libraries/APIs
TensorFlow, Keras, Pandas, PyTorch, Scikit-learn, OpenCV, GDAL/OGR, XGBoost, NumPy, SciPy
Tools
ChatGPT, Mathematica, GIS, Azure Machine Learning, Amazon SageMaker, Tableau
Languages
Python 3, Python, SQL, C++11, C++, Fortran, JavaScript
Frameworks
LlamaIndex, Sphinx Documentation Generator, Scrapy, Flask
Paradigms
App Development, Anomaly Detection, DevOps, Penetration Testing
Platforms
Jupyter Notebook, Linux, Docker, Amazon Web Services (AWS), Azure, AWS Lambda
Storage
JSON, Data Pipelines, MySQL, MongoDB
Other
Statistics, Machine Learning, Deep Learning, Computer Vision, Natural Language Processing (NLP), Data Science, Image Processing, Predictive Modeling, Convolutional Neural Networks (CNNs), Semantic Segmentation, Artificial Intelligence (AI), Computer Vision Algorithms, Data Analysis, Data Analytics, CSV, Statistical Modeling, Machine Vision, Neural Networks, AI Design, Software Engineering, Code Review, Source Code Review, Technical Hiring, Task Analysis, Modeling, Chatbots, OpenAI GPT-3 API, Generative Pre-trained Transformer 4 (GPT-4), Full-stack Development, Statistical Analysis, Predictive Analytics, Data Cleaning, Data Cleansing, LangChain, Random Forests, Supervised Machine Learning, OpenAI GPT-4 API, Programming, Large Language Models (LLMs), Scientific Computing, Scientific Data Analysis, Mathematical Modeling, Time Series Analysis, Geodetics, Physics, GeoPandas, FastAPI, Data Engineering, BERT, Medical Imaging, Object Detection, Algorithms, Object Tracking, Data Visualization, Time Series, Word2Vec, JSTransformers, Transformers, Cloud, Drone Photography & Videography, Cloud Services, Interviews, Interviewing, Team Management, APIs, Chatbot Conversation Design, Generative Pre-trained Transformers (GPT), Real-time Data, Point Cloud Data, Leadership, Remote Sensing, Advanced Physics, Numerical Modeling, Mathematics, Team Leadership, Voice Recognition, Speech Recognition, Generative Adversarial Networks (GANs), Machine Learning Operations (MLOps), Adversarial Machine Learning, Language Models, User Interface (UI), Software Architecture
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring