AI & software projects
Selected engagements framed as problem, approach and outcome — anonymised by default, named only where a client has agreed.
Claim document extraction for an insurance platform
Problem
An insurance claims platform received accident notification reports, odometer photos and vehicle ID plates as unstructured images attached to each claim. Staff retyped the same fields — odometer reading, VIN, accident narrative — by hand from every attachment before a claim could be triaged.
Approach
Built an event-driven pipeline that classifies each attachment, then routes it to the right extraction model: fast plate recognition for number plates, layout-aware OCR for odometer readings and VIN plates, running on serverless infrastructure that scales to zero between claims.
Outcome
Key-information extraction pulled VIN, odometer and ANR straight from photos.
Python · ONNX Runtime · YOLO · PaddleOCR · OpenCV · AWS Lambda · AWS Batch · Amazon SQS · Docker · FastAPI · PyTorch · scikit-learn · MLflow · Pydantic · Boto3
Repair quote extraction for automotive claims
Problem
Automotive repair quotes arrived as unstructured PDFs, and every labour line, part and total had to be re-keyed by hand before a claim could be compared or approved.
Approach
Built a document-AI pipeline that runs text detection ahead of a layout-aware model fine-tuned on real repair quotes, converting each PDF into a structured record of labour items, parts and totals through a set of coordinated services.
Outcome
Reached an 88.79% F1 score across 189 extracted entity types on 200+ real-world repair quotes, replacing manual re-keying with structured data.
Python · LayoutLMv3 · PaddleOCR · Flask · Redis · AWS EC2 · AWS S3 · AWS SNS · Supervisor · Docker · Pydantic · Jinja2 · s3fs · NLP · Computer Vision