Personal details
Nikolay M.
Summary
Enthusiastic and goal-oriented ML software development engineer with over 7 years experience in developing multiple business solutions using Python language. Having a degree in Computer Science, started his career path as a junior developer; demonstrated a proactive approach in career, skillset growth and problem-solving skills. Possesses expertise in Python, Golang, various Python/Golang frameworks, libraries and utilities. Has deep knowledge of Bash scripts and Linux. Has practice in building architecture for distributed applications: developed multiple interconnected services and microservices from scratch. Has hands-on experience working in an Agile environment with changing requirements and quick releases. Keeps focus on the result, always tries to choose reasonable and reliable solutions. Acts as a collaborative team player with excellent interpersonal skills.
Technical skills â‹… Main programming languages: Python, Golang, Rust, Bash â‹… Platforms: Linux Ubuntu, Linux CentOS â‹… Cloud Computing: AWS, GCP â‹… ML Platforms: AWS Sagemaker, Databricks â‹… Machine learning frameworks/libraries: TensorFlow, PyTorch, ONNX, OpenVINO â‹… Additional libraries: NumPy, pandas, pyspark, SciPy, sklearn, OpenCV, MLFlow; flask, sqlalchemy, fastAPI â‹… Neural networks: Fully connected, Convolutional, LSTM/RNN, GAN â‹… ML task fields: NLP, CV, Predictive analytics â‹… Data related: Feature engineering, ETL â‹… CI/CD: Jenkins, Github, Dockerhub â‹… Databases: PostgreSQL, MySQL, TimescaleDB â‹… Issue Tracking: Jira, Trello
Technical skills
Work Experience
H&M Migration models project: the migration of customer's models from Azure to GCP cloud. This includes a deep dive in the previous client’s infrastructure as well as designing a similar architecture (a better one) from scratch in GCP. To be able to migrate more models from client’s infrastructure, it was decided to prepare the generic architecture for ML models pipeline and then apply the same architecture with a minimum changes to the next customer’s ML model projects.
Responsibilities: · Team leading of ML Stream team; · Developing APIs from scratch using serverless GCP solutions; · Participating in design and architecture; · Designing and developing intercommunication between services; Technologies: · Python · Catboost, LightGBM, XGBoost, PyTorch · Azure, GCP, various Python SDKs for Azure/GCP · Docker, Kubernetes · pandas, numpy, scipy, sklearn
Developing and designing scenario generation and calendar evaluation; · Developing and improving the current code; · Improving stability, reproducibility and performance of the current solution; · Designing and developing business report and national calendar evaluations; · Introducing some fixes in the ML project; · Developing new 2 models and new features in Deep Learning part of ML project.
Kubernetes application catalog and Machine learning platform – Kibernetika.AI ,ex. KuberLab. Helps machine learning developers to create, build, train and deliver ML projects and improve speed of development. (https://kibernetika.ai) . Written in Go, Consists of a few key components: API, UI, cluster-management tool, application deployment service, service for automatic delivery of docker-based images to cluster, Datasets management and generic ML inference server (serving) Responsibilities: · Developing and designing microservice for user ML workloads; · Building and training neural networks using TensorFlow and PyTorch to use in high-order services (face/person recognition, face transformation, person tracking) · Developing key microservice: Dataset creation and management service, CLI for it and possibility to mount dataset or model as a filesystem (using Filesystem in User space Kernel protocol), developed in Go: https://github.com/kuberlab/pluk · Working on Generic Machine-learning inference server (aka ‘ml-serving’). It is capable of serving heterogeneous machine-learning models including Tensorflow, Intel OpenVINO, PyTorch, ONNX and is possible to build pipelines using any model and business logic. Written in Python
Technologies: · TensorFlow / PyTorch · Golang · Python · go-restful, Gorm, kubernetes, helm, net/rpc, x/oauth2, Google GCE API, AWS API, GitHub API, go-git.v4 · OpenVINO · OpenCV · NumPy, SciPy, Scikit-image, Sklearn