Senior ML Engineer

免费注册后解锁"你的匹配分 vs 高分画像"与个人技能缺口。

About the project:An AI platform that upgrades legacy CCTV with real-time detection, instant incident notifications, and decision-ready analyticsdesigned with GDPR and responsible AI at the core. Requirements. 4+ years of hands-on experience in computer vision, with a strong focus on embedded/edge device optimization; Deep knowledge of CNN-based object detection and face detection techniques; Expertise in model compression, quantization, and acceleration for embedded inference; Strong Python and C++ development skills; Experience with deployment tools: TensorRT, ONNX Runtime, TFLite, OpenVINO, or CoreML; Familiarity with embedded platforms (e.g., NVIDIA Jetson, Qualcomm Snapdragon, Raspberry Pi with Coral Edge TPU); Experience in profiling and optimizing models for latency, throughput, and memory efficiency. What you will do. Communication: Lead as part of a scrum team focused on delivering production-ready, optimized computer vision models. Promote clear, respectful, and constructive communication; Collaboration: Work closely with machine learning engineers, software developers, and hardware teams to ensure seamless integration of vision models into products; Documentation: Maintain high standards for documenting model training processes, optimization techniques, deployment pipelines, and performance benchmarks; Continuous improvement: Stay updated on state-of-the-art research in embedded computer vision and edge AI, bringing new ideas and technologies into the development process; Planning: Actively participate in sprint planning and provide accurate estimations, particularly for model optimization and deployment tasks. Model Development and Optimization: Design and optimize CNN-based object detection and face detection models; Implement advanced model optimization techniques, including pruning, quantization, and knowledge distillation; Develop models with a focus on efficient deployment on embedded or edge devices. Embedded Deployment: Use tools like TensorRT, ONNX Runtime, TFLite, OpenVINO, or CoreML for deployment; Work with embedded hardware platforms such as NVIDIA Jetson, Qualcomm Snapdragon, and Raspberry Pi with Coral Edge TPU. Performance Profiling: Profile models for latency, throughput, and memory efficiency; Optimize the inference performance to meet strict resource and timing constraints on target devices. Development Skills: Develop and maintain codebases in Python and C++ to support model integration and deployment pipelines. What you will get. Competitive salary and good compensation package; Exciting, challenging and stable startup projects with a modern stack; Corporate English course; Ability to practice English and communication skills through permanent interaction with clients from all over the world; Professional study compensation, online courses and certifications; Career development opportunity, semi-annual and annual salary review process; Necessary equipment to perform work tasks; VIP medical insurance or sports coverage; Informal and friendly atmosphere; The ability to focus on your work: a lack of bureaucracy and micromanagement; Flexible working hours (start your day between 8:00 and 11:30); Team buildings, corporate events; Paid vacation (18 working days) and sick leaves; Cozy offices in 2 cities ( Kyiv Lviv ) with electricity and Wi-Fi (Generator Starlink); Compensation for coworking (except for employees from Kyiv and Lviv); Corporate lunch + soft skills clubs; Unlimited work from home from anywhere in the world (remote); Geniusee has its own charity fund. Originally posted on Himalayas

关于项目：一个AI平台，通过实时检测、即时事件通知和可决策的分析来升级传统CCTV，并以GDPR和负责任AI为核心设计。要求。4年以上计算机视觉实操经验，重点关注嵌入式/边缘设备优化；深入了解基于CNN的目标检测和人脸检测技术；精通模型压缩、量化和加速以实现嵌入式推理；强大的Python和C++开发技能；具有部署工具经验：TensorRT、ONNX Runtime、TFLite、OpenVINO或CoreML；熟悉嵌入式平台（例如NVIDIA Jetson、Qualcomm Snapdragon、搭载Coral Edge TPU的Raspberry Pi）；具有分析和优化模型以提升延迟、吞吐量和内存效率的经验。你将做什么。沟通：作为scrum团队的一部分，主导交付生产就绪、优化的计算机视觉模型。倡导清晰、尊重和建设性的沟通；协作：与机器学习工程师、软件开发人员和硬件团队紧密合作，确保视觉模型无缝集成到产品中；文档：保持模型训练流程、优化技术、部署管线和性能基准文档的高标准；持续改进：持续关注嵌入式计算机视觉和边缘AI的最新研究，将新想法和新技术引入开发流程；规划：积极参与sprint规划并提供准确的估算，尤其是模型优化和部署任务。模型开发与优化：设计并优化基于CNN的目标检测和人脸检测模型；实施先进的模型优化技术，包括剪枝、量化和知识蒸馏；开发模型时重点关注在嵌入式或边缘设备上的高效部署。嵌入式部署：使用TensorRT、ONNX Runtime、TFLite、OpenVINO或CoreML等工具进行部署；使用嵌入式硬件平台，如NVIDIA Jetson、Qualcomm Snapdragon和搭载Coral Edge TPU的Raspberry Pi。性能分析：分析模型的延迟、吞吐量和内存效率；优化推理性能，以满足目标设备上严格的资源和时间约束。开发技能：使用Python和C++开发和维护代码库，以支持模型集成和部署管线。你将获得。有竞争力的薪资和良好的薪酬待遇；令人兴奋、富有挑战且稳定的初创项目，采用现代技术栈；企业英语课程；通过与来自世界各地的客户持续互动，练习英语和沟通技能；专业学习补偿、在线课程和认证；职业发展机会，半年度和年度薪资审查流程；执行工作任务所需的必要设备；VIP医疗保险或运动保障；非正式且友好的氛围；能够专注于你的工作：没有官僚主义和微观管理；灵活的工作时间（在8:00至11:30之间开始你的一天）；团队建设、企业活动；带薪休假（18个工作日）和病假；位于2个城市（Kyiv Lviv）的舒适办公室，配有电力和Wi-Fi（Generator Starlink）；联合办公空间补偿（Kyiv和Lviv的员工除外）；企业午餐+软技能俱乐部；可在世界任何地方无限期居家办公（远程）；Geniusee拥有自己的慈善基金。最初发布于Himalayas

⚑ 岗位信息收录自公开招聘页（Himalayas），以雇主官网为准