PROFESSIONAL EXPERIENCE
- Python programming, 5+ years of experience
- Main fields: Data science, Machine learning, Computer vision, Software engineering, ML engineering
TECHNICAL SKILLS
-
- Programming languages: Python (Keras, TensorFlow, PyTorch, Scikit-learn, OpenCV, Numpy, Numba, Pandas, Matplotlib, Django, Flask, Scrapy), JavaScript (React, VueJs), PHP
- Databases: SQL, MongoDB, ElasticSearch
- Web services: REST, RPC
- Team project management: Bitbucket, GitLab, Jira, Trello
- Operating Systems: Linux OS, Windows OS, Raspbian
- Software Design Methods: Agile
- Version Control Tools: GIT
- IOT: Raspberry Pi, on edge computers
- AWS services: Cloudwatch, Cognito, Kinesis, Kinesis Video streams, EC2, ECR, IAM, Fargate, Lambda, Rekognition, S3, Sagemaker, SQS, SNS
EDUCATION
- General education:
- BSc Software engineering, Kaunas University of Technology
PROJECTS (Agmis)
| Project: Retail Crime Predictor | 2019 11 - Present |
| Technologies used | Python (OpenCV, Numpy, Pillow, PyTorch, Tensorflow, Scikit-learn), YoloV4, Docker, AWS services (Lambda, SNS, SQS, ECR, S3, EC2, SageMaker), Convolutional Neural Networks, RabbitMQ, ElasticSearch |
| Role | Team Lead, ML Engineer |
| Project Description | The main objective - detect if a person is trying to buy an item with a permuted barcode. |
| Responsibilities and achievements | Solution architecture. Create data processing pipelines to process real-time scan events from the cashier with on edge devices as well as cloud infrastructure (AWS) for data storing, monitoring. Analyze data from stores, provide insights and apply these insights to create ML models. Develop custom classification model training pipelines and automation with SageMaker. Solution pilots are successfully running in multiple stores with plans to expand. |
| Project: Count people travelling by train | 2021-03 - 2021-03 |
| Technologies used | Python (OpenCV, Pandas, Numpy, Tensorflow, Matplotlib), AWS services (SageMaker, S3), Recurrent Neural Networks, YoloV4, Trackers |
| Role | ML Engineer |
| Project Description | The main objective - analyze how many people are traveling by train without a ticket |
| Responsibilities and achievements | Analyze the project and present possible solutions. Develop software to process video data from each of the train wagons and count people entering/leaving using object detection and optical flow trackers. |
| Project: Power Lines | 2020 07 - 2020 08 |
| Technologies used | Python (OpenCV, Pandas, Numpy, Numba, PyTorch, Matplotlib, Open3D), 3D classification |
| Role | ML Engineer, Data Analysis |
| Research Direction | Analyze LiDAR data and applicable solutions for further steps |
| Responsibilities and achievements | Analyze 3D data, prepare, visualize 3D data and processed results. Develop processing scripts to filter, crop and classify point cloud to provide initial results. |
| Project: Retail Shelf monitoring | 2020 04 - 2020 06 |
| Technologies used | Python (MXNet, OpenCV, PyTorch, Tensorflow, Numpy), Convolutional Neural Networks, AWS services (S3) |
| Role | ML Engineer |
| Project Description | The main objective - detect the absence of items on shelves. |
| Responsibilities and achievements | Analyze data and create ML models, logic development. Create ML models to classify each product in a shelf as well as empty spaces. Develop processing pipelines to classify data in real-time. Develop main logic to assign each item to a position according to the original planogram. |
| Project: Picklist assistant | 2020 03 - 2020 03 |
| Technologies used | Python (OpenCV, PyTorch, Numpy, MXNet), Convolutional Neural Networks, AWS services (S3, SageMaker) |
| Role | Data Analytic, ML Engineer |
| Project Description | Main objective - automate the selection of unwrapped products at the self-checkout counter |
| Responsibilities and achievements | Analyze data and create ML models to classify unwrapped products that are being weighted. Develop a real-time processing pipeline. Demonstrate the solution to the client. |
| Project: Welding monitoring | 2019 08 - 2019 10 |
| Technologies used | Python (Pandas, Matplotlib, Numpy), AWS services (SageMaker, Fargate, Lambda, Kinesis Video Streams, EC2, CloudFormation) |
| Role | Computer Vision Engineer |
| Project Description | The main objective - track how much time is needed for a welding action to be done in a video stream. |
| Responsibilities and achievements | Create a pipeline using AWS services to monitor welding, analyze and visualize data after the analysis period is completed. |
PROJECTS (Teltonika)
| Project: IoT devices management platform | 2018 05 - 2019 05 |
| Technologies used | PHP (Lumen, Laravel), VueJS, MySQL, GIT, Docker |
| Role | Software Engineer |
| Project Description | IoT devices management platform |
| Responsibilities and achievements | Write well documented and tested code. Developed new features, maintained existing code by fixing bugs and monitoring errors. |