2+ Year Computer Vision Engineer (OCR Pipeline)

About Us

We are building AI-first tools to automate and accelerate workflows in the construction industry. Our focus is on extracting insights from complex documents such as blueprints and invoices to streamline preconstruction and project management processes.

Role Overview

We are looking for a part-time AI Vision Engineer (OCR Specialist) with 2–4 years of hands-on experience in Computer Vision and OCR. The ideal candidate will have proven expertise in reading and interpreting construction blueprints, scanned PDFs, and invoices , and building analysis pipelines around them.

Responsibilities

Design and implement OCR workflows for parsing blueprint notes, legends, title blocks, and invoice documents.

Build scalable document analysis pipelines for text, tabular data, and mixed layout extraction.

Apply preprocessing techniques (deskewing, denoising, line / symbol isolation) for construction drawings.

Collaborate with the team to generate structured outputs (e.g., BOQ, invoice summaries, scope extraction).

Develop quality checks and error detection heuristics to ensure reliable extraction results.

Work closely with domain experts to map extracted data to construction-specific use cases.

Required Skills & Experience

2–4 years of industry experience in Computer Vision & OCR (Tesseract, EasyOCR, Donut, LayoutLM, etc.).

Strong background in document layout analysis (tables, legends, line items).

Experience with construction drawings and / or invoices (scanned PDFs, hybrid CAD exports).

Familiarity with text detection & recognition models (EAST, CRAFT, TrOCR).

Knowledge of preprocessing methods for noisy / rotated scans.

Proficiency in Python, OpenCV, and deep learning frameworks (PyTorch / TensorFlow).

Ability to deliver modular, production-ready code for integration with downstream systems.

Do not apply i f you have had live exposure to at least 2 years of Vision and OCR-related business problems. Job ID is Threetwoone5

Nice-to-Have

Exposure to blueprint object / symbol detection (YOLO, Faster R-CNN, UNet).

Experience in multimodal AI (Vision + NLP).

Construction domain understanding (MEP systems, takeoffs, invoice reconciliation).

Compenstation

Part-time / Flexible (remote).

15-20K INR / month. For the first 3 months | 30K from 4th month

Job ID - 200100