OCR Developer - Data Extraction from Mexican Payroll Receipts (Two Templates)

I'm looking for an experienced OCR/computer vision developer to build a data extraction solution for Mexican payroll receipts (recibos de nómina). Scope: - Extract all fields from the payroll receipt documents - Two fixed formats, both templates are consistent and do not change over time - Full field extraction from both versions (I'll provide real sample documents once we connect) - Output to structured data (format to be finalized, likely JSON/CSV into a downstream system) Accuracy & speed are critical: The extracted data feeds a downstream financial process, so extraction needs to be highly accurate and reliable, not "close enough." Processing also needs to be fast and efficient at volume. Volume: This needs to handle high throughput; the end use case involves processing receipts for a large employee base (tens of thousands), so the solution should be built for scale and batch processing, not just one-off documents. Note on the documents: These are specific Mexican payroll formats, so there's no off-the-shelf template for them; the solution will need to be built and configured specifically for these two layouts. Part of what I'm evaluating is how you'd approach that. What I'm looking for: - Proven OCR / document intelligence experience (please share relevant examples) - Comfort with fixed-template extraction and reliable structured output - Ability to handle volume/throughput - Availability to start quickly; this project is time-sensitive To apply, please include: - Relevant OCR/extraction projects you've done (especially structured documents like invoices, receipts, or forms) - Your proposed fixed price for a two-template extraction like this - Your estimated timeline - Your availability - The OCR engine/stack you'd use, and why. The documents are in Spanish; Spanish-language document experience is a plus but not required, since the fields are structured.