Research Analyst and Developer specialised in transforming complex datasets into actionable intelligence. My work sits at the intersection of Machine Learning, Computer Vision, and Data Engineering, with a focus on building end-to-end pipelines from research prototypes to scalable systems.
- Machine Learning and Computer Vision: YOLOv8/v11 Object Detection, PyTorch, ResNet Classification, Multi-stage Clustering (HDBSCAN), Cloud Vision API Benchmarking (AWS Rekognition, Azure, Google Vision).
- Data Engineering: Modular ETL Pipelines, REST API Aggregation (SERP, Alpha Vantage, Polygon.io), Async/Concurrent Processing, NAS Integration, Automated Data Updates.
- Financial Data and Analysis: Batch Backtesting Frameworks, Indicator Analysis (Tom DeMark), Market Cap Screeners, Portfolio Transaction Logs.
- Tools: Python, SQL, PowerBI, Google Colab, DGX (GPU Training), Jira, Confluence.
- Global Brand Discovery Architecture: Designed a multi-stage clustering and human-in-the-loop validation system for global brand identification across sports broadcasting.
- Computer Vision Pipeline: Built and benchmarked object detection pipelines using YOLOv8/v11 on DGX infrastructure, including SAM-assisted annotation workflows that reduced annotation time significantly.
- Cloud Vision API Benchmarking: Conducted systematic benchmarking between AWS Rekognition, Azure, Google Vision, and open-source solutions for production-ready brand detection.
- Technical Leadership: Authored 20+ internal guides and tutorials on LLMs, Deep Learning, Cluster Analysis, and ML workflows to support team capability building.
- UK House Price Index: Self-serving Python dashboard to visualise UK house price index by geography and time period.
- Datapopy: Python API wrapper for streamlined data extraction from data.police.uk.
- London Data Store: Module to search, filter, and extract data from the London Data Store.
- sodakit: Enhanced Python client for the Socrata Open Data API.
Currently focused on productionising ML pipelines for sports technology and expanding open-source contributions in computer vision.