About The Role
Build and support mining robotics systems with an emphasis on perception, navigation, and operational safety. You will improve autonomy reliability through rigorous dataset curation, evaluation, and QA practices across simulation and field logs.
What You Will Do
-
Develop and test autonomy features for mining robots (haulage, inspection, mapping) using ROS/ROS2 pipelines
-
Build evaluation harnesses for perception and planning, including scenario-based testing and regression tracking
-
Define labeling taxonomies and enforce annotation guidelines compliance for computer vision and sensor datasets
-
Run QA evaluation on labeled data, monitor inter-annotator agreement, and implement sampling plans
-
Track and document metrics (precision/recall, collision-risk proxies, localization drift, dataset shift)
-
Support RLHF-style human feedback and prompt evaluation for LLM assistants used in procedures and incident triage
-
Partner with vendors and internal teams to scale data labeling and validation throughput
Required Qualifications
-
Mid-Senior experience in robotics, autonomy, or applied ML
-
Hands-on knowledge of ROS/ROS2 and C++ and/or Python
-
Experience with SLAM, perception, computer vision, and/or sensor fusion
-
Familiarity with data labeling workflows, QA evaluation, and taxonomy design
-
Understanding of safety engineering concepts for industrial robotics
-
Strong communication skills for cross-functional remote collaboration
Workflows and Data
-
Multi-camera video, LiDAR point clouds, GNSS/IMU trajectories, telemetry time series, and event logs
-
Annotation audits, edge-case mining, evaluation set curation, and targeted re-labeling for model performance improvement
Workplace
This role is
Remote
and
Full-Time
. “Belo Horizonte” is not a location requirement; it is a search intent modifier.