Machine Learning Engineer in Constanţa

Software Engineering (regular) TensorFlow (regular) PyTorch (regular) Python (advanced) About AppsilonAppsilon is an ambitious and fast-growing software house and consultancy specializing in decision support systems and machine learning with Fortune 500 clients across the globe. We are a unique company driven by a mission to improve our society and environment. Some examples of our #data4good work include contributing to wildlife preservation in the National Parks of Gabon, building COVID-19 dashboards, and improving data science tools for Doctors Without Borders.In the machine learning space we specialize in computer vision, applying it in cases making impact on biodiversity and researching new approaches.We are also a global leader in R and Shiny, which are used by companies of all sizes to build analytical applications. When companies run into difficult problems or want to initiate large-scale enterprise projects, they come to Appsilon.Before you apply, please read our code of conduct.Every few months we start completely new projects and dive into a completely new world. One day we identify species of monkeys lurking behind trees of a rainforest, another day we analyze satellite images to help mitigate natural disasters, and then dive into the arctic ocean, helping researchers understand the changes in those ecosystems. Our projects are not only an opportunity to test our skills in difficult statistical, algorithmic, and technological problems but also an opportunity to learn about different research fields and, most importantly, contribute to their advancement. Some examples of our past projects:Open source wildlife detection app, built for usage in remote areasAnalysis of damage after natural disasters based on satellite images: https://demo.appsilon.ai/apps/building_damage_assessmentWe took 5/811 place in the Hakuna Ma-data competitionA research paper applying machine learning in ecological modellingYour Role as a Machine Learning EngineerRegular duties will include:Preparing dataCollecting, curating, possibly preprocessing a datasetEDA - exploratory data analysisUnderstanding and visualizing statistical properties and peculiarities of the dataset ModellingRunning and monitoring model’s training Investigating the model’s performance, identifying strong and weak pointsPipeline setup and improvementsMaking sure the process above is modular and reproducibleHandling meetings with the client / partnerSharing results, challenging assumptions, understanding the role of ML in their workflowHelpful skills and experienceHard skillsGreat Software Engineering backgroundExtensive Python knowledgeExperience with PyTorch or TensorflowExperience in data wranglingExperience with machine learning pipelines and experiment reproducibilitySoft SkillsTrained analytical thinkerAble to switch between hacker mentality of getting things to work and organized engineer adhering to basic principles when refactoring or building key pipeline elementsAble to abstract from technical issues and communicate also on high levelAt least B2 level of EnglishWhat’s in it for you?Salary 10000 - 16000 PLN + VAT on B2B contract26 days of paid holidays + an equivalent of public holidays in Poland, est. 11 days in 2021+5% of salary in Professional Development Budget to spend on activities that help you grow33 days (paid 80%) per year on B2B when on a sick leaveRemote work with flexible working hours adjusted to your time zone and family life.4 paid days per year to be used for training/conferences, events, or workshops for your professional developmentPrivate health care insurance (in Poland) (Polmed)Life insuranceFitProfit or FitSport membership card (in Poland)AskHenry – a personal assistant works great in large Polish cities, elsewhere limited to online supportProjects that have a real impact on the world. More on https://appsilon.com/data-for-good/Technologies and tools you will be usingAll sorts of Pythonic tools forData processing - pandas, numpy, Pillow, opencv, …Data visualisation - matplotlib, seaborn, plotly, …Modelling - PyTorch (and fast.ai), TensorFlow (and Keras), scipy, …Experiment tracking - Weights&Biases, Neptune, Domino, Tensorboard, …Dashboards - rather limited, but still maybe sometimes - streamlit, django, starlette, ...Occasionally R worldMostly to interface with Shiny dashboardsPossibly if data source/preprocessing was implemented in R by a clientCode tools (GitHub/GitLab, your favourite IDE, sometimes RStudio)General tools (bash/shell)Internal tools (Slack, Outline, Clickup, G Suite)What can you expect during the recruitment process?Screening call with Talent ManagerHome assignmentInterview with the ML Team Conversation with the CTO