Artsiom P.

Key Skills

Software

No software listed

Top Subject Matter

Maritime movement anomaly detection using AIS graph features

User churn/drop-off prediction and structured dataset preparation

Port call detection and unsupervised clustering of ports

Top Data Types

Computer Code Programming

Top Task Types

Classification

Freelancer Overview

OrbitMI — Data Scientist (Anomaly detection pipeline and automated delivery). Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include AWS ECS, Neo4j, and PySpark. Education includes Bachelor's degree, Harbour.Space (2025). AI-training focus includes data types such as Computer Code and Programming and labeling workflows including Classification, Clustering, and Computer Programming.

Labeling Experience

OrbitMI — Data Scientist (Anomaly detection pipeline and automated delivery)

Classification

Built an end-to-end anomaly detection ML pipeline for maritime movement using graph-based features and Isolation Forest. Implemented the solution as containerized services deployed on AWS ECS, including automated CI/CD for delivery and testing. Designed data processing and routing logic for large-scale AIS-derived datasets to support downstream model training and evaluation. • Engineered graph-based anomaly detection features and Isolation Forest workflow • Deployed containerized ML services on AWS ECS and maintained CI/CD pipelines • Implemented a routing API using Dijkstra graph search in Neo4j for navigation-related inference • Processed large-scale Big Data from AIS vessels using PySpark

2025 - Present

Stena Bulk — Data Scientist (Big Data ETL + EDA + clustering model training)

Engineered scalable Python ETL pipelines to process Big Data and detect port calls from 100M+ real-world records. Conducted exploratory data analysis for bunkering operations using Python to identify optimized fuel quantities and estimate potential savings. Applied advanced data modeling and unsupervised ML clustering to group unknown ports for operational insights. • Processed and analyzed 100M+ records using Dremio, SQL, and Airflow • Used Pandas-based exploratory analysis to optimize fuel quantity decisions • Performed unsupervised clustering of 800+ unknown ports using DBSCAN • Produced data models and analytics-ready outputs to support downstream analysis

2024 - 2025

Leagues of Code — Data Scientist Intern (Automated pipeline + supervised ML model training)

Classification

Developed automated data pipelines that transformed raw journal entries into structured datasets for analytics use. Trained a Random Forest model to predict user drop-off and applied the results to reduce client churn. Implemented automation scripts for targeted outreach to at-risk clients, supporting data-driven decisioning. • Built ETL-style pipeline for transforming unstructured journal entries into structured data • Trained Random Forest for churn/drop-off prediction and assessed impact on client churn • Automated targeted outreach workflow using Python scripts based on model outputs • Bridged ML outputs to business operations and analytics-ready datasets

2023 - 2024

Tera — Python Developer (Automated data extraction for dataset preparation)

Extracted and parsed eCommerce product data from Amazon and normalized datasets using Python. Implemented asynchronous programming to improve data preparation performance for large product catalogs. Focused on preparing structured, analytics-ready data for downstream usage. • Wrote Python scripts for extraction and parsing of eCommerce data • Used asynchronous programming to normalize and prepare datasets • Cleaned and structured 60,000+ product records for further processing • Improved reliability and readiness of prepared data for analytics pipelines

2021 - 2022

Education

Work History