Senior Machine Learning Engineer
Description
The ASUS Robotics & AI Center is seeking a Senior Machine Learning Engineer to join our global research and development team. This is an applied engineering role, not a research position. It centers on leading the design, delivery, and long-term performance of computer vision models that turn real-world imagery into accurate 3D and semantic understanding for our next-generation autonomous platforms.
We are looking for a seasoned engineer who brings deep technical expertise, sound engineering judgment, and a proven track record of delivering machine learning products that customers depend on in complex, real-world environments. The ideal candidate is a self-directed contributor who can set performance standards, drive technical decisions, mentor others, and operate effectively across a collaborative, multidisciplinary team.
Roles and Responsibilities
- Lead the design, development, and deployment of machine learning systems for computer vision and 3D perception tasks.
- Own the performance of deployed models, setting accuracy, latency, and reliability targets and making sure they are met.
- Optimize training and inference pipelines for cost and throughput on cloud infrastructure.
- Build pipelines that combine imagery with 3D and spatial data.
- Establish practices for monitoring, retraining, and updating models in production.
- Evaluate new techniques in deep learning, object detection, segmentation, 3D reconstruction, and visual tracking, and bring the ones that hold up into production.
- Own the data strategy for vision models: labeling guidelines, annotation tooling and QA, class imbalance and rare classes, and active learning.
- Drive technical decisions and provide guidance on architecture, tooling, and best practices across the team.
- Design and execute experiments, including benchmark datasets and real-world field testing, to validate model performance.
- Maintain and improve datasets, pipelines, and tools to support efficient model training and deployment.
- Collaborate with cross-functional teams, including robotics, systems, and software engineers, to deliver production-ready solutions.
Requirements
Required Qualifications
- Bachelor’s degree or higher in computer science or robotics.
- 5+ years of industry experience developing and deploying machine learning models for commercial applications, such as computer vision, with a track record of delivering products. Time in academic or research roles does not count toward this requirement.
- A proven track record of leading machine learning systems into production, where products or customers depend on them, and maintaining them at scale.
- Experience setting and meeting performance requirements for deployed models, such as accuracy, latency, throughput, or uptime targets.
- Experience with detection and/or segmentation models on real-world imagery where data is imperfect, imbalanced, or scarce.
- Proficiency in Python and deep learning frameworks such as PyTorch and/or JAX. Hands-on experience training models such as VLMs, ViTs, and/or CNNs.
- Familiarity with classical computer vision techniques (e.g., OpenCV).
- Strong experience with MLOps practices, such as model versioning, monitoring, and CI/CD for ML.
- Exceptionally strong problem-solving skills and the ability to lead technical work in a fast-paced, multidisciplinary environment.
- Understanding of software development best practices, including coding standards, code reviews, source control management, and test automation.
- Excellent written and verbal communication skills in English.
Preferred Qualifications
- A master’s degree or PhD, or a degree specialization in machine learning, computer vision, geometric computing, or 3D vision.
- A solid grounding in 3D geometry, multi-view reconstruction, and camera models.
- Experience with robotics, autonomous systems, or real-time perception applications.
- Experience with camera pose estimation, structure from motion, SLAM, or 3D reconstruction methods such as Gaussian splatting, NeRF, or mesh processing.
- Experience with C++ or GPU programming (e.g., CUDA, Triton).