Skip to main content
Subhajit Bhar - IDP Engineer

Subhajit Bhar

AI Document Processing | OCR Automation | RAG & AI Agents

I design automated document pipelines that turn messy files into clean, structured data using Python, AI APIs (OpenAI, Anthropic), and Microsoft VBA. My work focuses on scalable and reliable document processing, and building maintainable data extraction workflows that hold up when your documents change.

Clients hire me when their document automation starts failing—works for some files, breaks on edge cases, fails silently on new layouts. I strengthen business logic, boost extraction quality, and implement review checkpoints (confidence scoring, LLM-assisted validation, citation tracking, human-in-the-loop validation) so your team can trust the output.

Recent projects:

  • Built Python/FastAPI web apps for document processing—covering lab-report PDF analysis, multi-step review and validation, confidence scoring, document ingestion from email, OCR, and cloud deployment.
  • Debugged and stabilized VBA workflows, taking over and rescuing projects mid-way.
  • Delivered custom Python scripts, APIs, dashboards, and VBA scripts tailored to clients’ needs.

I’m an Expert-Vetted freelancer on Upwork with 100% Job Success, $50K+ earned, and 10+ completed contracts.

Wondering if automation is right for your documents? Book a free 30-minute call for an honest assessment.

Book a Free 30-min Call →

Recent

Certificate of Analysis Data Extraction: A Production Guide

·8 mins
A certificate of analysis (CoA) is one of the most information-dense documents in regulated industries. It carries test results, method references, accreditation details, chain-of-custody information, and the laboratory’s sign-off — all in a format designed for human reading, not machine parsing.

Contract Data Extraction: Pulling Structured Data from Legal Documents

·8 mins
Contracts are the hardest document type to extract data from reliably. Invoices have a predictable structure. Lab reports have defined fields. Contracts are natural language documents, and the information you need — key dates, party names, payment terms, renewal clauses, termination conditions — can appear anywhere, phrased in many different ways, across documents that range from two pages to two hundred.