Machine Learning Engineer
Developed end-to-end ML pipelines for Document AI, specializing in NLP tasks and multimodal models for PDF files.Developed multiple proof-of-concept solutions for document classification, layout classification, entity extraction. This includes an encoder-based transformer model that consistently achieves over 97% accuracy in the production pipeline.Integrated and prompt-engineered a GenAI model with Langchain and OpenAI API for special cases of entity extraction.Deployed several model on AWS Sagemaker Inference endpoints.Deployed a centralized logging environment(CloudWatch, S3) which gather log data from docker containers and AWS resources.