I will build a document processing and ocr pipeline on AWS textract and lambda


About this gig
Have piles of PDFs, scans or letters that someone types into a system by hand? I build pipelines that read documents automatically and send the data where you need it.
What I do:
- OCR with AWS Textract: text, forms and tables
- Serverless processing on AWS Lambda, S3, SQS and SNS
- Throttling and buffering to stay within Textract limits
- Retries and dead-letter queues, so no document gets lost
- Delivery of results to your database, API or a Rails app
Why me: I owned an OCR pipeline in production at a German company: Lambda, Textract and SNS/SQS. I designed a buffering layer to throttle Textract, built an hourly S3 exchange with a data science classifier and automated intake of documents from Zendesk.
Please message me before ordering with a few sample documents and what data you need from them.
Get to know Dmitrii T
Backend Engineer
- FromUkraine
- Member sinceNov 2026
Languages
Russian, Ukrainian, English
Other Software Development Services I Offer
FAQ
Which documents can you process?
PDFs, scans and photos of printed documents. Handwriting works worse, so I check it on your samples first.
Do I need an AWS account?
Yes, the pipeline runs in your account, so you own it and pay AWS directly.
Can you integrate it with my system?
Yes, through your API, a database or a Rails app.
