Machine Learning Engineer
hace 11 horas
Stanford-le-Hope
Role: Machine Learning Engineer (Robotics) About AssetCool AssetCool is addressing one of the most pressing challenges in the global energy transition - grid congestion. By combining cutting-edge grid robotics with breakthrough materials science, our technology transforms the physical performance of the grid in situ, increasing current-carrying capacity by up to 30%, at just 5% of the cost and in a tenth of the time compared to conventional approaches. In addition to enhancing capacity, our solutions tackle other grid constraints including corrosion, electrical noise, and enable the collection of high-value data to build accurate, predictive digital twins. This technology is proving transformative and is being adopted at an accelerating pace, with deployments across Canada, Slovenia, the UK, and the USA, and a strong global deployment pipeline. AssetCool is one of the fastest growing start-ups in Europe, with plans to more than double, again, in size over the next six months. Backed by international investors from the US, Europe, and the UK, we are proud to be one of the most innovation-focused companies in the UK energy sector. We’re looking for ambitious, mission-driven individuals who are excited to work hard alongside their colleagues to build and deploy world-leading technology that tackles some of the planet’s most urgent infrastructure challenges. Role Overview We are looking for a Robotics Machine Learning Engineer to join our software team and help deliver ambitious ML solutions that power real-world robotic inspection and automation. This is a hands-on, applied role where you’ll build multi-modal machine learning models that operate in challenging outdoor environments, helping take ideas from prototype to field-tested deployment. You’ll be involved in the full ML lifecycle: building datasets, training and evaluating modern vision models, running experiments, and supporting deployment to edge devices. Demonstrable experience is essential - this could be through a previous role, research, open-source contributions, or strong personal/university-level projects. Candidates at all levels are encouraged to apply. Key Responsibilities • Develop and improve machine learning pipelines for robotic inspection tasks, including: segmentation; detection; classification; object tracking., • Build, train, and evaluate deep learning models using modern architectures such as:, • YOLO model family architectures for object detection and image segmentation., • Transformer-based object detection models (e.g., DETR / RT-DETR style approaches)., • Segmentation models such as Segment Anything Model (SAM), • Multimodal Vision Language Models (e.g. PaliGemma, Qwen2.5-VL, SmolVLM)., • Work with real-world industrial datasets (variable lighting, clutter, motion blur, weather effects) and help define clear annotation and labelling standards., • Support annotation workflows and dataset management using platforms such as Roboflow., • Prototype solutions quickly, moving from idea → training → evaluation using clear metrics (IoU, mAP, precision/recall, F1, etc.)., • Train ML systems using both real and synthetic datasets, including:, • generating synthetic training data for rare events and edge cases., • applying domain randomisation techniques to generalise training outcomes., • evaluating sim-to-real transfer performance., • Perform statistical and numerical analysis to evaluate model performance, failure modes, and robustness in field conditions., • Collaborate with robotics, embedded, and software teams to ensure ML systems are testable, deployable, and reliable on real hardware., • Communicate results clearly through short technical write-ups, sprint demos, and experiment logs. Required Skills & Experience • Strong Python programming skills, including experience with ML training and data workflows., • Experience with Jupyter Notebooks., • Demonstrable hands-on experience developing and training vision models., • Architectural fluency and a deep understanding of state-of-the-art architectures, including transformers and CNNs with the ability to select and adapt models based on specific constraints and applications., • Experience with deep learning frameworks such as PyTorch (preferred) or TensorFlow., • Understanding of key ML/CV concepts:, • dataset splits, overfitting, evaluation metrics, augmentation, and performance trade-offs., • Comfortable working with image processing and common computer vision libraries (e.g., OpenCV, NumPy, Pillow, SciPy, Matplotlib)., • Good engineering habits: Git-based workflows, virtual environment usage, clean code, and reproducible experiments., • Strong problem-solving ability, able to iterate quickly and learn new tools and techniques. Preferred Skills & Experience • Experience with Roboflow for dataset management, augmentation, annotation workflows, and/or model training., • Comfortable using Jupyter Notebooks for rapid experimentation, dataset exploration, and model evaluation., • Experience with “cutting-edge” vision model ecosystems and customisation (e.g. prompt-driven workflows, foundation segmentation models, hyperparameter tuning, adapters)., • Experience with synthetic data generation and simulators for perception training and testing., • Familiarity with robotics simulation tooling such as NVIDIA Isaac Sim, MuJoCo, Gazebo (or similar simulation environments)., • Familiarity with deployment constraints and optimisation for edge inference (e.g. NVIDIA DeepStream, or similar tooling), • Experience with Docker or container-based development for reproducible training and deployment environments., • Understanding of practical ML engineering workflows: experiment tracking, model/version management, dataset provenance, and regression testing, • Basic C/C++ knowledge (helpful for integrating models into production robotic systems)., • Keen interest in robotics, real-time perception, and the latest machine learning and computer vision research, with an emphasis on practical deployment outside the lab. Why Join AssetCool? • Work on groundbreaking technology revolutionizing the power grid., • Fast-growing startup with significant industry traction and large-scale projects., • Collaborative, innovation-driven team with exciting career growth opportunities., • Competitive salary, and flexible work options Research shows that some candidates may hesitate to apply unless they meet every listed requirement. If this role excites you but your experience doesn’t align perfectly with every qualification, we still encourage you to apply. You might be exactly who we're looking for, either for this role or another opportunity within our team. We’re proud to be an equal opportunities employer and welcome applications from people of all backgrounds. We’re committed to building an inclusive, supportive workplace where everyone can thrive, regardless of age, disability, gender identity, marital or civil partnership status, pregnancy or maternity, race, religion or belief, sex, or sexual orientation. If you have a disability or any specific requirements and need adjustments at any stage of the recruitment process, just let us know and we’ll do our best to accommodate your needs. We may close this vacancy early if we receive a high volume of applications. We encourage you to apply as soon as possible to avoid missing out.
Página 6