Wrote a document processing API server (POST to https://api.aryn.cloud/v1/document/partition) in Python using Ray and FastAPI that turns PDFs into images, does object detection, and segments the PDF text, images, and tables in order to enrich data for searching across documents with LLMs and RAG. Tables are turned into JSON. Configured the server on AWS using Terraform. UI at https://app.aryn.ai.

Embedded CNTK & ONNX model training & inference C++ library in a NoSQL database.

Developed parameter checking code for embedded API’s.