Machine Learning Engineer – Computer Vision & Document AI [Job ID: 2068915224]

Degree Names

Description

We are looking for a Machine Learning Engineer to lead the development of AI models that understand construction drawings, extract relevant data, and generate accurate material estimates.

Responsibilities

• Develop and deploy deep learning models to:
• Extract annotations, text, dimensions using OCR
• Interpret legends, tables, and specification sections in drawings
• Preprocess architectural/engineering drawings (PDF, DWG, images) for model consumption
• Build hybrid models that combine CV, OCR, NLP, and rule-based estimation logic
• Design and maintain scalable pipelines for training, inference, and feedback loops
• Work closely with domain experts (estimators, civil engineers) to translate business rules into algorithms
• Continuously improve model accuracy and robustness using real-world data
• Document processes, datasets, and model evaluation techniques

Requirements

• 5 years of experience in machine learning, especially computer vision and document AI
• Proficient in Python and ML frameworks like PyTorch or TensorFlow
• Strong experience with OCR (Tesseract, EasyOCR, or cloud OCR APIs)
• Hands-on knowledge of object detection (YOLO, Detectron2, Faster R-CNN)
• Experience with PDF and CAD parsing (e.g. PDFMiner, ezdxf, AutoCAD APIs)
• Knowledge of preprocessing scanned documents and dealing with noise/low quality data
• Solid understanding of model evaluation, data labeling, and annotation tools

Nice to Have:

• Experience with layout understanding models (Donut, LayoutLM, PubLayNet)
• Knowledge of civil and construction work.
• Exposure to pricing/estimation logic or ERP systems
• Experience deploying ML models in production (Docker, FastAPI, etc.)

Required Skills