Senior Machine Learning Engineer in Constanţa

We are taking part in the development of a healthcare startup. The main purpose is to monitor patients in the hospital.  The main feature is an application for 24-hour video surveillance. Now we're developing an AI-based system to prevent falls from the bed. It's an important part of nurses' job to make sure no one falls, and this feature is very important. The client is one of the world’s largest manufacturers of appliances and electronics. Stack: TypeScript/JavaScript, Node.js, React, MySQL, Kubernetes, Docker, Elastic Kubernetes Service, Terraform, Helm, AWS, Twilio Video SDK For AI: JS, OpenCV, Tensorflow, Tensorflow JS, PoseNet, Coco SSD, SNPE, C++, Python Architecture: Service-oriented application that utilizes multiple edge devices — smart cameras Team composition: 5SE, 2QA, 2ML, AQA, UI/UX designer, DevOps Client-side team structure: Product Owner, 3SE Project highlights: The application is in production Development practices on the project include Code review, Build-in quality, Unit-testing and integration testing, End-to-end automated testing, CI, Release on demand Working in a distributed multinational team in a different time zone. The approximate work schedule is from 11:00 AM to 7:00 PM CEST Project stage: The project has passed the MVP phase. Application is deployed to prod. Active development continues in parallel with production support. Workflow:  2-week sprint. The input of the process will be sprint backlog, output — functionality increment. The focus here is on the predictable delivery of defined pieces of functionality Continuous exploration process. The input of this process is some custom feature, improvement idea, or need. Output is a team backlog planned through sprints Scope of tasks and ownership: Solution Development & Prototyping: Work with business, clarify user needs, create solution hypotheses and evaluate them Solution Prototyping: Develop prototypes using web runtime (JS) to check solution feasibility Collaborate with developers to define solution/model integration into existing implementation Viability validation: Work with UI/UX designers and Product Owner to make sure the prototype solves business needs Feasibility validation: Define metrics to measure, set up experiments, and develop benchmarks or collaborate with developers to find out the feasibility of proposed approaches Model Research & Training: Find existing and train models that solve the business task. Fine-tune them to increase accuracy Models Preparation: Convert models for execution in Web runtime using TF.js, ONNX or in C++ using Snapdragon Neural Processing Engine Papers Research: Find, compile, and design approaches to solve non-trivial tasks using the latest published research papers Datasets Preparation: Define requirements for evaluation/training data, work with business to get data, preprocess and annotate it using existing state-of-the-art approaches, or define requirements for annotation Performance Optimization: Continuously optimize algorithms and models for efficient (near-)real-time execution, considering resource constraints Models Evaluation & Testing: Design approach and test algorithms and models performance under diverse scenarios and edge cases What You’ll need: +5 years of relevant experience Strong expertise in computer vision techniques, including object detection, tracking, and image/video analysis Software Development experience using React/Node.js, Python, or other programming language is a must Knowledge of C++ is a big plus Familiarity with libraries, s.e. TensorFlow JS, OpenCV, PoseNet, Coco SSD Experience with deep learning frameworks and architectures, including CNNs, RNNs, and attention mechanisms Solid understanding of data preprocessing, augmentation, and feature extraction techniques for video data Ability to optimize algorithms for real-time or low-latency execution, considering hardware constraints Deep knowledge of models training, optimization, and investigating best options for a case Excellent problem-solving skills, critical thinking, and attention to detail Level of English Upper-Intermediate and above Our benefits: No micromanagement Freedom to engage in decision-making and implementation Ability to work in a team of professionals (the ratio of middle and above specialists 80/20) Participation in the development of high-quality products Direct communication with clients on a partnership level Professional development opportunities ($600 education budget, well-managed processes, communities, internal library) Health insurance $600 extra for the health care, sports, or mental health 20 paid working days off and 10 days of sick leave Opportunity to work remotely Soulful team buildings and corporate events Join us and be among those who care!