alkuz

Data Science team code review - both Data Science and Back-End parts. Closely worked with customer management team to explain Data Science process. Developed serverless backend for chatbot to collect data (AWS API Gateway, Lambda, SQS, RDS, CloudWatch). Developed PoC application for processing changes in documents to notify users (Python Multiprocessing, regular expressions). Developed serverless backend to integrate NLP model with customer cloud and implemented integration with Outlook server (AWS Lambda, Step Functions, S3, API Gateway, SQS, O365). Supervised and developed manually models for emails classification, discussed metrics, both models and data quality with the customer, prepared presentations and documentation in Confluence (SKLearn, Pandas, AWS Comprehend). Building baseline object detection models based on AWS Rekognition service. Supervised and developed manually different computer vision models for classification, segmentation, and object detection tasks (Tensorflow, Keras, TFOD, PyTorch). Developed web interface for customer to discuss data labeling process (AWS EC2, Streamlit). Building data labeling pipelines based on crowdsource and customer teams (AWS Mechanical Turk, boto3, JupyterLab). Improved seismology data quality for geological research in Oil-Gaz industry using Computer Vision techniques (Tensorflow, Keras, Matplotlib, JupyterLab). Financial transactions analysis for fraud prediction - building pipeline for data quality analysis (PowerBI, Pandas, Python Multiprocessing).

Dataset collection: parsing images from open sources and OpenImages Dataset v4. Data preparation: creating preprocessing pipelines with Scikit-learn, Open-CV, Pandas, Numpy, in addition I used HDF5 and Tensorflow Records for data packaging. Training and validation of computer vision models: OCR, object detector (TFOD), image classifier (Tensorflow, Keras, also using transfer learning and pre-trained models as backbones for custom architectures). Development of the microservice architecture - RestAPI, business logic and ML; Microservices deployment via Docker compose; ML model deployment via Tensorflow Serving and Docker; Connection to video broadcast over the network, running machine learning model predictions, description of the objects interaction logic in frames. Implementation of the RestAPI on Flask. Implementation of multi-threaded calculations in Python. DS Task management: decomposition of tasks for other team members, managing of daily meeting with team, retrospective and planning at the end of sprint, backlog grooming.

Research of intelligent control systems of Aviation Engines. Research stability of Engine’s control systems. Laboratory test bench development for Engine’s work simulation.

Aviation Engines models development in Matlab Simulink. Control Systems development and simulation in Matlab tools. Aviation Engine identification via Neural Networks.