Oleksandr Shyshkin

Key Skills

Software

Appen

Labelbox

Sama

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Evaluation Rating

Prompt Response Writing SFT

Segmentation

Translation Localization

Freelancer Overview

With a unique multidisciplinary background in Clinical Psychology, Computer Science, and Physics, I bring a highly analytical and rigorous approach to AI data training and evaluation. Over the past year, I have gained extensive hands-on experience as a freelance AI trainer with industry leaders such as TELUS International, Appen, and Sama, focusing on complex LLM evaluation and computer vision annotation. My work has involved detailed text analysis, including quality and safety assessments, factual accuracy verification, and sentiment analysis. In computer vision, I am proficient in tasks ranging from bounding boxes and object tracking to precise semantic segmentation. My academic role as a research assistant at the University of Geneva has allowed me to apply these skills directly, leveraging LLMs to analyze data for therapeutic approaches. This, combined with my native proficiency in four languages (English, French, Ukrainian, Portuguese), enables me to handle nuanced, multilingual data with exceptional accuracy. I am adept at quickly mastering complex project guidelines and am committed to delivering the highest quality data to improve AI systems.

Labeling Experience

Clause-Level Relationship Extraction from Financial Reports

AppenTextEntity Ner ClassificationRelationship

This is a project for a FinTech client aimed at training an NLP model to understand financial documents. I am tasked with performing multi-level Named Entity Recognition on dense financial reports to identify entities like companies, executives, and monetary values. The core task was relationship extraction, annotating the specific legal and financial relationships between these entities at the clause level, according to an extensive and highly technical set of guidelines.

This is a project for a FinTech client aimed at training an NLP model to understand financial documents. I am tasked with performing multi-level Named Entity Recognition on dense financial reports to identify entities like companies, executives, and monetary values. The core task was relationship extraction, annotating the specific legal and financial relationships between these entities at the clause level, according to an extensive and highly technical set of guidelines.

2025

3D Point Cloud Annotation for Autonomous Delivery Robots

Sama3D SensorCuboidObject Detection

Worked on a project for a robotics company developing autonomous urban delivery vehicles. The task involved annotating 3D point cloud data from LiDAR sensors and fusing it with corresponding 2D video footage. I was responsible for accurately drawing 3D cuboids around dynamic objects (pedestrians, vehicles) and performing semantic segmentation on the point cloud data to identify terrain types. This required focus and the ability to work within a complex 3D annotation environment.

Worked on a project for a robotics company developing autonomous urban delivery vehicles. The task involved annotating 3D point cloud data from LiDAR sensors and fusing it with corresponding 2D video footage. I was responsible for accurately drawing 3D cuboids around dynamic objects (pedestrians, vehicles) and performing semantic segmentation on the point cloud data to identify terrain types. This required focus and the ability to work within a complex 3D annotation environment.

2025 - 2025

Medical Image Segmentation for AI Diagnostics

SamaImagePolygonPoint Key Point

For a client in the psychiatric AI research field, I performed precise segmentation on structural brain MRIs. My task was to accurately delineate key brain regions associated with mood and memory disorders, such as the hippocampus, amygdala, and prefrontal cortex. This highly detailed annotation was used to train an AI model to identify potential biomarkers for psychiatric conditions. The project required exceptional attention to detail and the ability to work with complex neuroanatomical guidelines. I was chosen for the project due to my ongoing Psychology studies which have a large emphasis on neuroanatomy.

For a client in the psychiatric AI research field, I performed precise segmentation on structural brain MRIs. My task was to accurately delineate key brain regions associated with mood and memory disorders, such as the hippocampus, amygdala, and prefrontal cortex. This highly detailed annotation was used to train an AI model to identify potential biomarkers for psychiatric conditions. The project required exceptional attention to detail and the ability to work with complex neuroanatomical guidelines. I was chosen for the project due to my ongoing Psychology studies which have a large emphasis on neuroanatomy.

2024 - 2025

Search Quality Rater for a Major Search Engine

TelusTextClassificationEvaluation Rating

Acted as a Search Quality Rater, evaluating the relevance and quality of web pages in response to user search queries. This involved a deep analysis of user intent and meticulous application of a 200 page set of guidelines to grade results. The role required strong analytical skills and the ability to make consistent judgments on content quality and user satisfaction.

Acted as a Search Quality Rater, evaluating the relevance and quality of web pages in response to user search queries. This involved a deep analysis of user intent and meticulous application of a 200 page set of guidelines to grade results. The role required strong analytical skills and the ability to make consistent judgments on content quality and user satisfaction.

2024 - 2025

Social Media Feed and Ad Relevance Annotation

AppenTextClassificationEmotion Recognition

Worked on a long-term project for a global social media platform to improve the relevance of its user feeds and advertisements. My responsibilities included classifying posts based on content and tone (sentiment analysis), evaluating the relevance of ads to user interests, and flagging inappropriate or low-quality content. This project required a high degree of cultural and linguistic nuance.

Worked on a long-term project for a global social media platform to improve the relevance of its user feeds and advertisements. My responsibilities included classifying posts based on content and tone (sentiment analysis), evaluating the relevance of ads to user interests, and flagging inappropriate or low-quality content. This project required a high degree of cultural and linguistic nuance.

2024 - 2025

Show more

Education

Work History