Work Experience
Seek AI
New York City, New York
NLP Researcher / Engineer
September 2022 - Present
- Developed a SoTA NL to SQL pipeline utilizing knowledge graphs, RAG techniques, and multiple models
- Built a pipeline monitoring platform with Kafka to evaluate SQL outputs using a custom metrics suite
- Finetuned LLM’s from Meta and Mistral with multi-GPU VMs from GCP on various code generation tasks
- Researched and implemented findings in CoT prompting, recursive error correction, and agentic behavior
- Assisted in a $10.5 million seed round and a projected $1m ARR through direct customer engagement
Viewpoint Software
Portland, Oregon
Data Science Intern
June 2020 - September 2020
- Imported raw data from 2,000,000 construction projects, anonymized and cleaned it to build usable datasets
- Modeled data with Python and created projections based on client needs, integrating within Viewpoint’s ERP
Software Engineering Intern
June 2019 - September 2019
- Modernized a 20-year-old C codebase to improve Viewpoint’s ERP functionality, reliability, and readability
- Implemented new payroll features, directly affecting 8,000 global clients from the construction industry
Ice Events
Edinburgh, Scotland
Co-Founder and Director
April 2019 - June 2022
- Developed profitable brand, business, and marketing strategies for 1,800-person entertainment venues
- Secured mutually beneficial promotional deals with local businesses and societies annually worth £50,000
Skills and Abilities
Progaming Languages:
Software:
- PyTorch
- Huggingface
- Docker
- Github
- Linux
- Snowflake
- Weaviate
- Postgres
- Kafka
- GCP
Projects
Freelance Photographer
June 2016 - Present
- Captured professional level landscapes, portraits, and architectural images, with 6+ years of
experience
- Edited images using Adobe Photoshop and Adobe Lightroom for a consistent Portfolio appearance
- Partnered with multiple international clients from Africa, Europe, and North America
Generative Adversarial Networks
Machine Learning.pdf
- Trained a DCGAN on 237 screens from The Legend of Zelda, generating 100 unique and playable levels
- Overcame data diversity issues with Illumination by implementing Covariance Matrix Adaptation MAP-Elites
Neural Network Energy Consumption
Dissertation.pdf
- Benchmarked four generations of GPU at a range of power limits using TensorFlow’s ResNet-50 model
- Extrapolated results to find that optimal GPU efficiency occurs at 70% of the manufacturer listed TDP
Education
University of Edinburgh
September 2017 - June 2021
BSc (Honors) Artificial Intelligence
UK 2:1 / US 3.6 GPA
Courses:
- Algorithmic Data Structures
- Computation and Logic
- Data Analysis
- Linear Algebra
- Probability and Statistics
- Software Testing
- Applied Machine Learning
- Computer Security
- Discrete Math
- Multivariable Calculus
- Quantum Computing
- Vision and Robotics
Societies:
- Computer Society
- Southern African Society
- Economics Society
- Third Culture Society