Hi there We are South and our client is looking for a
Machine Learning Engineer
Note To Applicants
- Eligibility: This position is open to candidates residing in Latin America.
- Application Language: Please submit your CV in English.
Applications submitted in other languages will not be considered.
- Professional Presentation: We encourage you to showcase your professional experience by including a Loom video in the application form.
While this is optional, candidates who provide a video presentation will be given priority.
We are looking for a Machine Learning Engineer with strong expertise in computer vision and object detection to tackle a specialized challenge: detecting walls and identifying rooms in architectural blueprints or pre-construction plans.
This role is ideal for someone who is hands-on, detail-oriented, and passionate about applying advanced ML techniques to solve real-world problems.
Key Responsibilities
- Develop and fine-tune ML models for detecting walls in architectural blueprints.
- Design and implement preprocessing pipelines for handling blueprint files (images/PDFs).
- Optimize model inference using ONNX Runtime for production-ready deployment.
- Build and maintain an API server to serve ML inference (preferably using BentoML).
- Collaborate with stakeholders to define requirements, deliver milestones, and document solutions clearly.
Must-Have Requirements
- Strong hands-on experience with PyTorch for model development and training.
- Expertise with Ultralytics (YOLOv8) for object detection tasks.
- Proficiency in OpenCV, Pillow (PIL), and NumPy for preprocessing and image manipulation.
- Experience with ONNX / ONNX Runtime for optimized inference.
- Solid knowledge of BentoML (preferred) or FastAPI for serving ML APIs.
- Familiarity with Pydantic for schema validation.
- Proven track record of deploying ML models into production.
- Strong problem-solving mindset and ability to adapt ML tools to specialized use cases.
- Clear communication and documentation skills.
Nice-to-Have
- Experience working with PyMuPDF for parsing PDF-based architectural plans.
- Background in architectural/engineering data or prior work with blueprint analysis.
- Knowledge of clustering/grouping methods for room identification tasks.
- Familiarity with MLOps practices (monitoring, scaling, CI/CD for ML).
Schedule
: Monday to Friday - Full-time
Compensation
: USD salary
Location
: 100% remote
If this opportunity sounds good to you,
send us your resume