Principal Software Engineer – Dataset and Annotation Quality - Physical AI, Autonomy & Data Platform

Caterpillar Inc.•Irving, TX

1d•$159,120 - $258,570

About The Position

Caterpillar is seeking a highly experienced Principal Software Engineer to define and lead the system design of annotation and dataset quality systems supporting machine learning and AI for physical systems and autonomy. The role designs how raw sensor data becomes high-quality labeled datasets, ensuring training data is accurate, consistent, and aligned with model needs. As the technical authority for annotation and dataset quality, this engineer defines system designs, standards, and engineering patterns for teams to implement and scale. The role is hands-on and addresses labeling consistency, ontology design, quality validation, human-in-the-loop workflows, and continuous dataset improvement in an environment where requirements evolve with model performance and real-world behavior. The position works across data quality architecture, data-centric AI feedback loops, software engineering, and technical leadership. It partners with machine learning, autonomy, product, data engineering, annotation engineering, platform engineering, and DevOps teams to turn objectives into technical roadmaps and deliver reliable, production-grade systems. The position is based in Irving, Texas (Dallas); domestic relocation assistance and visa sponsorship are available.

Requirements

10+ years of software engineering experience, including significant experience in principal-, staff-, or lead-level technical leadership roles.
Expert-level proficiency in Python and/or Java.
Expertise in computer vision, sensor-based systems, or ML-driven products.
Strong experience designing annotation platforms, labeling systems, or data quality frameworks.
Strong understanding of data-centric AI and training data lifecycle management.
Experience defining label ontologies, annotation guidelines, and quality standards.
Experience building or scaling human-in-the-loop systems.
Strong understanding of data quality metrics and their impact on model performance.
Experience connecting model failures to data issues and driving improvements.
Proven ability to lead architectural direction across multiple teams.
Deep expertise in system design, distributed systems, and large-scale cloud-native architectures.
Strong experience designing and implementing streaming and batch data processing systems at enterprise scale.
Hands-on experience with AWS cloud services and architecture patterns.
Experience leading technical strategy across multiple engineering teams in agile environments.
Demonstrated ability to work effectively in highly ambiguous or rapidly evolving technical environments.
Comfort with experimentation, prototyping, and iterative architecture refinement.
Ability to make sound technical decisions with incomplete information and evolving constraints.
Expert knowledge of software development tools and activities, SDLC, software integration, software product design and architecture, technical product aspects, and software testing.
Extensive knowledge of decision-making processes and techniques, effective communication, and software integration processes.

Nice To Haves

Bachelor’s or Master’s degree in Computer Science, Software Engineering, Computer Engineering, or a related field.
Experience with Azure or GCP cloud services and architecture patterns.
Experience supporting AI/ML, autonomous systems, computer vision, robotics, or advanced analytics platforms.
Experience with multimodal data annotation strategies involving LiDAR, images, and telemetry.
Familiarity with annotation tools or platforms such as CVAT, Labelbox, or SuperAnnotate.
Experience with auto-labeling, active learning, or hybrid annotation systems.
Experience leading or influencing annotation operations at scale, internally or through vendors.

Responsibilities

Define data quality frameworks, including validation rules, scoring models, and acceptance criteria.
Design scalable systems for label validation, consistency checking, and anomaly detection.
Define data quality metrics, including consistency, agreement, completeness, and fitness for model use.
Establish dataset readiness, production acceptance, auditability, and dataset governance standards for machine learning systems.
Design approaches for ongoing data quality monitoring and improvement.
Define how model performance informs annotation prioritization and strategy.
Design systems for active learning, error mining, and targeted data improvement.
Establish mechanisms to improve datasets continuously based on real-world model behavior.
Provide design direction to annotation and dataset quality engineering teams; review detailed designs for alignment with system architecture.
Partner with machine learning, autonomy, and product teams to ensure data quality drives system performance, and serve as the organization's technical authority for annotation and dataset quality.
Lead software development using Python and/or Java in enterprise-scale environments.
Champion SDLC practices, including CI/CD, automated testing, infrastructure as code, code quality, release management, and operational maturity.
Establish engineering practices that support reliability, observability, maintainability, and platform stability.
Participate in hands-on development, prototyping, troubleshooting, and performance tuning of critical platform components.
Drive modernization initiatives and continuous improvement of engineering processes and platform capabilities.
Promote iterative engineering that balances rapid innovation with production-grade discipline.
Collaborate with Product Management, Data Engineering, Annotation Engineering, Platform Engineering, and DevOps teams.
Translate business and platform objectives into actionable technical roadmaps for scrum teams.
Provide technical leadership across agile teams and maintain alignment with architectural strategy.
Facilitate design reviews, sprint planning, backlog refinement, dependency management, and engineering governance.
Promote collaboration, accountability, adaptability, and engineering rigor.