Senior AI Perception Engineer
- Arizona - Remote - USA; Michigan - Remote - USA; California - Remote - USA · Remote
- Full-time
- Posted
Skills
Job description
Job Description:
Senior AI Perception Engineer
Location: United States, Remote
Type: Full-time
Travel: Approximately 10%, including quarterly onsite meetings at a Nextpower California office and visits to solar farms or test sites
Reports to: Senior Director, Robotics & Autonomy
About Nextpower
Nextpower is a global leader in intelligent solar tracker and software solutions, optimizing solar power plant performance worldwide. Our Robotics and Autonomy team develops and deploys robotic inspection systems for live and under-construction solar sites. Construction environments can be busy, crowded, and constantly changing, with active equipment and people requiring careful coordination. Once a site is operating, it may be remote and quiet, with robots working for long periods in open, sun-exposed conditions and little or no human presence. Our systems combine mobility, perception, embedded computing, autonomy, and cloud intelligence to increase inspection capacity, improve safety, and scale operations.
We are seeking a Senior AI Perception Engineer to provide hands-on technical leadership for computer vision and perception systems used in robotic inspection. You will shape the technical direction, build and deploy models, and own the software and data workflows that take perception from development through field validation and release. You will help the team turn robot-collected data and human inspection expertise into reliable, measurable improvements in automated inspection.
Role Overview
This is a senior individual-contributor role for an engineer who can move between model behavior, software architecture, and real system performance. You will lead the design and deployment of perception capabilities across robot-side nodes and cloud-assisted workflows, applying rigorous statistical methods to data, model evaluation, and fault assessment. You will work with mission management to close the loop between model results and targeted data collection, and you will set a high technical bar for testing, releases, and operational performance.
You will work closely with the autonomy and reliability leads and the broader engineering team to deliver a working system and shape the inspection roadmap. You will visit test and operating sites to observe workflows, collect representative data, and investigate system behavior across busy construction environments and remote, sun-exposed operating sites. This role combines deep computer vision and machine learning expertise with strong software engineering, system-level judgment, and practical robotics experience.
Key Responsibilities
Computer Vision and Model Development
· Set technical direction for perception and inspection models, including object detection, recognition, tracking, classification, segmentation, and fault detection.
· Design, train, evaluate, and improve models for inspection tasks, balancing accuracy, confidence, latency, compute, and operational requirements.
· Define dataset strategy and data quality practices; work with inspection specialists to identify labeling needs, data gaps, edge cases, and opportunities to learn from human decisions and demonstrations.
· Build and maintain repeatable training and evaluation pipelines, including dataset, experiment, and model versioning.
· Apply probability, statistics, and rigorous experimental methods to model development, uncertainty estimation, anomaly detection, and fault assessment, including statistical distance methods such as Mahalanobis distance where appropriate.
· Lead the design and evaluation of sensor-fusion approaches using vision, lidar, GPS, IMU, and, where applicable, multispectral data.
Benchmarking and Testing
· Define benchmarks, evaluation methodology, and performance criteria across fault classes, equipment, sites, environmental conditions, and data-capture variations.
· Analyze false positives, false negatives, uncertainty, and other model errors; identify root causes and drive prioritized improvements.
· Establish repeatable validation using field data, simulation, and log replay to test expected behavior and difficult edge cases.
· Ensure model and software changes are tested as part of the release process, with documented results, limitations, and release evidence.
Deployment and Field Validation
· Own the technical architecture and lifecycle of perception and inspection nodes running on the robot, including interfaces, performance, health reporting, deployment, and integration with the wider system.
· Design robot-to-cloud workflows that use APIs and distributed compute for inference or analysis when appropriate, accounting for connectivity, latency, bandwidth, cost, and degraded-operation behavior.
· Integrate model results with mission management so detected faults, low-confidence results, or data gaps can trigger targeted collection of additional images, viewpoints, or sensor measurements.
· Optimize and deploy models on embedded or GPU-based hardware, profile performance on target systems, and resolve issues found in integration and field testing.
· Visit construction and operating sites to observe inspection workflows, collect representative data, and evaluate perception performance in busy, changing environments and remote, sun-exposed sites.
Software Delivery and Team Execution
· Establish and reinforce strong software engineering practices for maintainable, tested, version-controlled software and released models.
· Lead technical design and code reviews, provide hands-on problem solving, mentor teammates, and build shared standards that improve the quality and consistency of the team’s work.
· Create and review pull requests, address feedback, and ensure automated tests and release checks cover integrated model and software changes.
· Break down technical work into trackable Jira tasks, communicate progress and risks, and present technical decisions and results clearly to engineering and cross-functional teams.
· Partner with the autonomy and reliability leads to align perception interfaces, system behavior, release readiness, and the technical roadmap.
· Document model behavior, architecture, test methods, evaluation results, operational limitations, and deployment requirements.
Minimum Qualifications
· Bachelor’s degree in computer science, robotics, electrical engineering, data science, or a related field; or equivalent practical experience.
· 5 or more years of experience developing and deploying computer vision or machine learning systems, including systems used in production, field, or other real-world operating environments.
· Strong programming skills in Python and substantial experience with a deep learning framework such as PyTorch.
· Demonstrated experience designing and evaluating models for object detection, recognition, tracking, classification, segmentation, and visual fault detection or inspection.
· Experience training, benchmarking, optimizing, and deploying models on embedded or GPU-based hardware.
· Familiarity with experiment tracking and dataset or model management tools such as MLflow.
· Experience designing or evaluating sensor-fusion systems that combine vision, lidar, GPS, IMU, or multispectral inputs.
· Strong working knowledge of probability, statistics, benchmarking, and model evaluation, with the ability to use quantitative evidence to guide engineering decisions.
· Experience designing maintainable software, writing automated tests, debugging deployed systems, and testing integrated software or model changes for release.
· Experience using Git-based workflows, including creating pull requests, participating in code reviews, and addressing review feedback.
· Experience using cloud-based APIs or distributed compute workflows to scale model inference, analysis, or training while coordinating with edge systems.
· Experience developing software for mobile robots, including ROS or ROS2 nodes, sensor interfaces, and deployed robot workflows.
· Experience with simulation, software-in-the-loop (SIL), hardware-in-the-loop (HIL), test-driven development, or other rigorous methods for validating robotics and machine learning systems.
· Experience building learning pipelines from human demonstrations, observations, or inspection decisions.
· Experience providing technical leadership through design reviews, code reviews, mentoring, shared engineering practices, and hands-on problem solving.
· Experience documenting technical work, breaking it into trackable tasks in Jira or a similar system, communicating progress and risks, and presenting results clearly.
· Coachable and collaborative, with the ability to take technical feedback, apply it, and contribute effectively on a fast-moving engineering team.
· Ability and willingness to test systems on physical robots and travel to test or operating sites.
Preferred Qualifications
· Experience with autonomy concepts such as motion planning, localization, and state estimation, and how perception outputs support those capabilities.
· Experience with semi-supervised learning that combines simulated and real-world data.
· Experience creating or using photorealistic simulation environments and log-replay systems to generate, reproduce, and evaluate perception scenarios.
· Experience using AI-assisted development tools, such as Cursor, Codex, or similar coding agents, to accelerate implementation, testing, and documentation while independently validating the output and maintaining ownership of code quality.
· Experience with solar equipment, industrial inspection, construction, or other visual quality applications.
Join Us
Join a fast-moving team building technology that will change how solar sites are inspected. Nextpower is an industry leader, and our robotics team is putting capable robots to work in real environments, from active construction sites to remote solar farms. You’ll see your technical leadership and engineering work move from data and models into deployed systems, where better perception can improve inspection quality, expand operating capacity, and help people focus on the hardest decisions.
We’re excited to bring on an experienced, hands-on engineer who is curious, coachable, and eager to keep growing. You’ll have opportunities to deepen your technical expertise, take on broader ownership, and help shape the team’s direction as the work and your capabilities develop.
Nextpower is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.
At Nextpower, we are driving the global energy transition with an integrated clean energy technology platform that combines intelligent structural, electrical, and digital solutions for utility-scale power plants. Our comprehensive portfolio enables faster project delivery, higher performance, and greater reliability, helping our customers capture the full value of solar power. Our talented worldwide teams are redefining how solar power plants are designed, built, and operated every day with smart technology, data-driven insights, and advanced automation. Together, we’re building the foundation for the world’s next generation of clean energy infrastructure.
Nextpower is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.
We are Nextpower