We are seeking to build a stand alone OCR Engine to process 120,000,000 public records. The repository will contain mostly legal documents pertaining to Real Estate. The applicant should have experience in working with legal documents or public records at large. More importantly have a deep understanding of the following skills:
+Building Trainable Models in OCR
+Natural Language Processing
+Use of Mixed Precision for Inference
The extraction of key pair values specifically;
1. Parcel ID/APN numbers
2. Legal Descriptions
4. Plantiff & Defedant (Parties of a court case)
5. Street Address
6. Plat Book & Page & Lot & Block
7. Case Number
8. Judges Name
9. Outcome of case (Disposition)
10. Type of case (Foreclosure/Tax Lien) etc
11. Lender Name
12. (Amount of mortgage / Amount of Foreclosure) etc
13. Any and all tables that exist in the PDF
Any additional key pair values will be adjusted to our scope and allow you to adjust your proposal on a case by case basis.
Following the inference layers added and trained using tensor or like kind machine learning technology, you will move the pdfs to cooresponding s3 buckets for efficient indexing.
26 freelancer đang chào giá trung bình $911 cho công việc này
I'm an AI Engineer, specialized in Computer Vision and CNN, i made a lot of projects and have a lot of certifications in deep learning and computer vision. As I'm an engineer, i care about performance and runtime.