Laurence Cullen
Verified Expert in Engineering
Machine Learning Developer
Laurence is a machine learning engineer and founder of venture-backed startup Vanellus, developing accelerated physics simulations. Laurence combines extensive experience executing and deploying R&D heavy machine learning projects with six years of industrial software engineering. His machine learning experience ranges from LLMs and computer vision to audio synthesis and cutting-edge approaches to accelerate numerical programming problems.
Portfolio
Experience
Availability
Preferred Environment
Machine Learning, Python, Go, TensorFlow, Keras, ChatGPT, Artificial Intelligence (AI), OpenAI GPT-4 API, Natural Language Processing (NLP), Computer Vision
The most amazing...
...models I have trained run physics simulations radically faster, help victims of abuse in the Indian Armed Forces, and translate ancient Sumerian into English.
Work Experience
Founder
Vanellus
- Developed a 2D fluid simulation solver from scratch for engineers, accelerated with cutting-edge machine learning research.
- Raised £400,000 of venture capital funding to progress business.
- Hired and trained academic mathematicians into industry-grade software engineers.
- Wrote grants for R&D funding projects for UK innovation agency.
Machine Learning Engineer
Self-employed
- Handled text classification of legal documents using BERT finetuning.
- Developed a neural singing system implementing the paper Arxiv.org/abs/1704.03809 in Keras.
- Created a Rasa chatbot to help users navigate questions around NDAs.
Machine Learning Engineer
Sensity
- Assembled from scratch a video scraping (Python and youtube-dl), labeling (Vue), and dataset building pipeline for deep fake videos found in the wild.
- Trained CNN models in PyTorch to accurately identify deep fake videos.
- Built the company's MVP comprising a React web app deployed with Firebase and dependent on several async detector services able to identify a variety of media manipulations.
- Productionized the deep fake detection model combining a preprocessing pipeline coordinated with a Pub/Sub architecture and deployed on GCP.
- Built a customer-exposed API in Go using the Go kit toolkit to allow external customers to interface with our detector systems.
- Contributed to the majority of features and infrastructure, leading to the company's successful seed round.
- Developed a named entity recognition pipeline using spaCy to extract the names of people mentioned in video titles.
Data Analyst and Engineer
Owlstone Medical
- Reduced the run time of test suites from 30 minutes to 5 minutes.
- Developed statistical and machine learning methods for detecting and classifying chemical weapons using Keras and Python.
- Audited core Python libraries' performance and drove large-scale improvements using high-performance libraries like Numba.
- Performed data exploration and analysis of the discriminatory power of company spectrometers in telling chemicals apart.
Software Engineer
Fetch.ai
- Integrated LiDAR and high-precision GPS sensors onto a drone flight platform using Arduino and Python.
- Built a simulation environment in Python to test drone navigation strategies.
- Processed imagery collected during drone surveys of agricultural land to provide actionable insights for farmers.
- Prototyped novel cryptocurrency architectures and tested performance under variable network conditions.
Experience
Sumerian to English Translation System
https://github.com/Laurence-Cullen/cuneiformThis involved creating a tokenizing model from the text using Google's SentencePiece tool and building word and word fragment embeddings from the entire corpus to use as a pre-trained embedding layer for the translation model with Facebook's FastText encoding tool.
Finally, I built a sequence-to-sequence model using an encoder-decoder LSTM neural network in Keras and trained it on the fraction of the Sumerian corpus for which translations exist. Some of the better experiments have achieved BLEU scores of 12.8, giving pretty sensible results for a good fraction of untranslated sentences.
Medical Ontology Alignment
https://github.com/Laurence-Cullen/ontology_alignmentI used the recently released Google sentence embedding model BERT to build embeddings of SNOMED and HPO terms and matched them based on cosine similarity. The results were surprisingly effective, and in most cases, the system could correctly translate a term from one ontology to another.
Bellingcat Tech Fellowship
https://github.com/Laurence-Cullen/bellingcat-hackathonI developed a tool to make the army tribunal documents published by the Indian government searchable with keyword search, semantic search, and reverse document search. This was achieved by combining web scraping of tribunal documents with the embedding of these documents using OpenAI API calls and a vector database to calculate distances between document embeddings. I deployed it into a web application with Google Cloud Run as a FastAPI server wrapped in a Docker container.
Skills
Languages
Python, JavaScript, Go, Regex, C
Libraries/APIs
OpenCV, Keras, TensorFlow, Pandas, Vue, Beautiful Soup
Paradigms
Data Science
Platforms
Unix, Linux, Docker, Google Cloud Platform (GCP), Amazon Web Services (AWS)
Other
Machine Learning, Convolutional Neural Networks (CNN), Natural Language Processing (NLP), Deep Learning, GPT, Computational Fluid Dynamics (CFD), Software Architecture, APIs, OpenAI, Physics, Neural Networks, LSTM Networks, Artificial Intelligence (AI), ChatGPT, OpenAI GPT-4 API, Computer Vision, Numerical Analysis, Venture Capital, FastAPI, Technical Leadership, Chatbots
Education
Bachelor of Science Degree in Astrophysics
University of Exeter - Exeter, UK
Certifications
Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization
Coursera
Structuring Machine Learning Projects
Coursera
Neural Networks and Deep Learning
Coursera
Machine Learning
Coursera
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