Tóm tắt công việc
You want to be an important member of a strong team of other makers who take pride in their work and support each other. As a Senior AI Engineer, you will join the AI-Powered Takeoff Team in Vietnam, collaborating closely with U.S. stakeholders and coordinating with external vendors to build an AI-driven product that transforms technical drawings (PDFs, CAD files, and blueprints) into actionable data, directly improving product performance and business efficiency. That's year one. If career growth is important to you, we not only know how to help you with that, it's what we love to do:
PROJECT DESCRIPTION
AI-Powered Takeoff is an intelligent construction TakeOff platform designed to automate material quantity takeoffs from architectural and engineering drawings. By leveraging computer vision, machine learning, and the system extracts, measures, and categorizes building components directly from digital blueprints PDF, significantly reducing manual effort, errors, and turnaround time.
The objective of AI-Powered Takeoff is to:
Automate quantity detection from drawings
Assist with measurements (length, area, count, volume)
Organize extracted quantities in structured formats
Support revision comparison
Reduce manual tracing effort
WHAT YOU'LL BE DOING
At Simpson Strong-Tie we value the individual, and the creativity that individuals can bring to the workplace. You join this company for a career, not a job!
You will join the AI Engineering team within an innovative manufacturing and construction technology environment, collaborating closely with the US team to design and develop AI services that enhance product performance and deliver measurable business impact.
You thrive in a cross-functional environment, working closely with product managers, AI engineers, subject matter experts (SMEs),
data analysts, product engineers, and business stakeholders to deliver high-impact solutions.
Design, develop, and maintain scalable end-to-end machine learning pipelines for automated vision processing
Build systems for symbol detection, drawing interpretation, floor plan parsing, and quantity extraction from construction documents (PDFs, CAD exports, blueprints).
Develop and fine-tune object detection and segmentation models to identify construction elements (walls, doors, windows, fixtures, framing, hardware components, etc.).
Implement OCR/YOLO and document layout analysis pipelines to extract annotations, dimensions, and labels from technical drawings.
Apply advanced ML techniques including transformer-based vision models and multi-label classification.
Conduct model evaluation, benchmarking, hyperparameter optimization, and performance tuning.
You will also be responsible for creating and optimizing core modules of the MLOps platform, from the computation layer to the orchestrator layer.
Collaborate with
software engineers to integrate ML models into production applications.
Manage and oversee the AI vendor's deliverables, including reviewing training approaches and datasets, and leading testing/UAT to evaluate and validate model accuracy and performance against business requirements.
If you can do everything listed above, you've got what it takes. Perhaps some of the following would be helpful too:
2+ years in Software Development (preferably construction tech, engineering, or document automation)
3+ years in AI/ML or Computer Vision roles
Strong hands-on Python experience for image processing and CV pipelines
ML & Computer Vision Expertise
Design and maintain scalable end-to-end ML pipelines for automated vision processing (symbol detection, drawing interpretation, floor plan parsing, quantity extraction from PDFs/CAD/blueprints).
Solid experience with OpenCV, Torchvision, YOLO, Detectron2, SAM, or similar.
Proven background in Object Detection, Image Segmentation, OCR, and Document Layout Analysis (e.g., PaddleOCR, Tesseract, LayoutLM, Donut).
Experience analyzing large sets of technical drawings and applying data augmentation for engineering documents.
Image classification, multi-label detection, semantic & instance segmentation
Transformer-based vision models (ViT, DETR, LayoutLM)
Model evaluation (mAP, IoU, F1), hyperparameter tuning, model optimization
Strong experience with PyTorch, TensorFlow, Scikit-learn, Pandas, Albumentations, etc.
Experience with Azure/Azure ML, AWS, or GCP
Other requirements (nice to have)
Deploy and optimize ML systems (low-latency inference, retraining pipelines, drift monitoring)
Model serving tools (TorchServe, ONNX, Triton, TF Serving)
Comfortable working in Linux environments
Ability to communicate technical outputs to non-technical stakeholders
Strong written and spoken English
Bonus: Experience with BIM (IFC/Revit), PDF parsing, GIS/spatial data, or construction/proptech domains.