Core Responsibilities / Key Responsibilities
1. Reusable Algorithm Module & Platform Development
· Lead the development of core algorithm modules decoupled from specific ships or eVTOL platforms, including high-precision SOH/RUL prediction model, early thermal runaway and fault warning model, real-time carbon footprint calculation engine, etc.
Lead the development of a core algorithm module library decoupled from specific ship or eVTOL platforms, including high-precision SOH/RUL prediction models, early thermal runaway and fault warning models, and a real-time carbon footprint accounting engine.
· Build a modular and microservices-based algorithm architecture, clearly define standardized interfaces (APIs), and support algorithm capabilities as independent services that can be quickly integrated and called by ship cloud platforms, eVTOL operation and maintenance platforms, and third-party systems.
Design a modular, microservices-based algorithm architecture with clearly defined standard interfaces (APIs), enabling algorithmic capabilities to be seamlessly integrated and invoked as independent services by ship cloud platforms, eVOTL operation platforms, and third-party systems.
· Establish an algorithm continuous training and iterative optimization pipeline, using multi-scenario data to continuously improve algorithm accuracy and generalization performance.
Establish a continuous training (CT) and iterative optimization pipeline for algorithms, leveraging multi-scenario data to consistently enhance accuracy and generalization.
2. Cross-Domain Application & Algorithm Migration
· Responsible for adapting, optimizing and validating the core algorithm module for different conditions such as ships (high vibration, high salt fog, long endurance) and eVTOL (high power, high safety, lightweight).
Adapt, optimize, and validate core algorithm modules for distinct operating conditions such as those in ships (high vibration, high salinity, long endurance) and eVTOLs (high power, high safety, lightweight requirements).
· Research and apply domain-adaptive and transfer learning techniques to improve algorithm reuse efficiency in different scenarios and significantly reduce redundant development costs.
Research and implement domain adaptation and transfer learning techniques to improve algorithm reusability across varied applications and significantly reduce redevelopment costs.
3.
Lead algorithm deployment in a cloud-edge-end collaborative architecture, designing integrated technical solutions that combine complex model training in the cloud with lightweight real-time inference at the edge.
· Optimize the computing efficiency, energy consumption, and reliability of edge-side algorithms to meet the needs of autonomous and intelligent maintenance of ships and aircraft in weak or no network environments.
Optimize the computational efficiency, energy consumption, and reliability of edge-side algorithms to enable autonomous intelligent operations in weak- or no-network environments for marine and aerial applications.
4. Data Strategy & Innovative Research / Data Strategy & Innovative Research
· Develop multi-scenario battery data standards and integration rules, build a unified data lake, and provide continuous and high-quality data support for algorithm research and development.
Formulate multi-scenario battery data standards and fusion protocols, and build a unified data lake to provide sustained, high-quality data for algorithm development.
· Track and introduce cutting-edge technologies (such as deep learning, physical information neural network PINN, digital twins), explore the feasibility of the next generation of intelligent maintenance algorithms and promote their implementation.
Monitor and integrate cutting-edge technologies (e.g., deep learning, Physics-Informed Neural Networks, digital twins) to explore and implement next-generation intelligent operation algorithms.
Requirements / Qualifications
Education & Experience / Qualifications & Experience
· Computer Science, Artificial Intelligence, Data Science, Control Theory, etc. related professional master's or doctoral degrees.
Master's or Ph.D. in Computer Science, Artificial Intelligence, Data Science, Control Theory, or a related field.
· More than 5 years of industrial algorithm architecture or development experience, with successful implementation and deployment of reusable algorithm modules or platform.
5+ years of industrial algorithm architecture or development experience, with a proven record of building and deploying reusable algorithm modules or platforms.
· Must have cross-domain algorithm deployment experience, covering at least two fields among new energy ships, electric vehicles, eVTOL/drones, energy storage, and industrial Internet.
Cross-domain algorithm deployment experience is essential, covering at least two of the following: new energy vessels, electric vehicles, eVTOL/UAV, energy storage, industrial IoT.
Professional skills / Technical Competencies
· Proficient in machine learning and deep learning, with solid theoretical and practical accumulations in time series prediction, anomaly detection, transfer learning, etc. Familiar with Python and PyTorch/TensorFlow frameworks,
Proficient in machine learning/deep learning with substantial theoretical and practical experience in time series forecasting, anomaly detection, and transfer learning. Skilled in Python and frameworks such as PyTorch/TensorFlow.
· Have deep software engineering and system architecture skills, familiar with microservices, Docker containerization, API design, and MLOps processes.
Strong software engineering and system architecture skills, familiar with microservices, Docker containerization, API design, and MLOps processes.
Knowledge of cloud-edge-end technology stacks; experience with edge computing frameworks (e.g., TensorRT, OpenVINO, AWS Greengrass) is a plus.
· Have excellent system thinking, able to transform complex business requirements into executable algorithm problems and technical routes.
Excellent systems thinking ability, capable of translating complex business needs into executable algorithm problems and technical roadmaps.
Comprehensive skills / Soft Skills
Outstanding capability in abstraction and architecture design, able to identify common requirements across applications and design universal technical solutions.
· Excellent cross-team communication skills, able to efficiently collaborate with battery R&D, embedded development, cloud platform and business teams to advance projects.
Excellent cross-team communication skills, with the ability to collaborate effectively with battery R&D, embedded software, cloud platform, and business teams.
· Have strong product and business orientation, focus on the commercial value and user experience of algorithm implementation.
Strong product and business orientation, focused on the commercial value and user experience delivered by algorithms.
· Able to communicate and write technical documents fluently in Chinese and English.
Proficient in both Chinese and English for work communication and technical documentation.
· Candidates with battery Prognostics and Health Management (PHM) related project experience are preferred.