For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

LoginCreate account
Invite to Job
I
Ivan Kaseko

Ivan Kaseko

AI Data Specialist / Quality Auditor

Poland flagPoznan, Poland
$30.00/hrExpertScale AIRemotasksOneforma

Key Skills

Software

Scale AIScale AI
RemotasksRemotasks
OneFormaOneForma
Internal/Proprietary Tooling
Other

Top Subject Matter

RLHF tasks
conversational AI
preference modeling

Top Data Types

TextText
ImageImage
Computer Code ProgrammingComputer Code Programming

Top Task Types

RLHFRLHF
ClassificationClassification
Red TeamingRed Teaming
Computer Programming/CodingComputer Programming/Coding
Prompt + Response Writing (SFT)Prompt + Response Writing (SFT)
Evaluation/RatingEvaluation/Rating
Data CollectionData Collection
Object DetectionObject Detection
Question AnsweringQuestion Answering
Text SummarizationText Summarization

Freelancer Overview

AI Data Specialist / Quality Auditor. Brings 4+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal and Proprietary Tooling. Education includes Bachelor of Science, WSB Merito University (2024) and Diploma, ZS im. J. i W. Zamoyskich (2024). AI-training focus includes data types such as Text and labeling workflows including RLHF, Classification, and Evaluation.

ExpertEnglishUkrainianPolishRussian

Labeling Experience

AI Data Specialist / Quality Auditor

TextRLHF
As an AI Data Specialist / Quality Auditor at Outlier.ai, I performed large-scale data labeling for RLHF tasks across multiple languages. My duties included preference ranking, helpfulness assessment, and safety evaluation on text datasets, maintaining high productivity and 95%+ accuracy. I collaborated with data science teams to refine annotation guidelines and ensure seamless integration into ML pipelines. • Annotated text data in Polish, Russian, Ukrainian, and English • Conducted quality audits and improved team annotation guidelines • Structured data into JSON/Airtable formats with strict schema validation • Processed over 500+ annotations per week while maintaining quality standards.

As an AI Data Specialist / Quality Auditor at Outlier.ai, I performed large-scale data labeling for RLHF tasks across multiple languages. My duties included preference ranking, helpfulness assessment, and safety evaluation on text datasets, maintaining high productivity and 95%+ accuracy. I collaborated with data science teams to refine annotation guidelines and ensure seamless integration into ML pipelines. • Annotated text data in Polish, Russian, Ukrainian, and English • Conducted quality audits and improved team annotation guidelines • Structured data into JSON/Airtable formats with strict schema validation • Processed over 500+ annotations per week while maintaining quality standards.

2024 - Present

RLHF Preference Ranking Dataset Project

Text
In the RLHF Preference Ranking Dataset project, I labeled over 5,000 response pairs for RLHF training, evaluating according to helpfulness, accuracy, safety, and coherence. My work included documenting preferences for model interpretability and maintaining high label consistency, even for complex multi-turn conversations. The project contributed to robust model evaluation and alignment with human preferences. • Performed detailed RLHF preference labeling for AI assistant evaluation • Evaluated responses with multi-criteria analysis and documentation • Managed consistency for diverse and challenging conversational data • Facilitated model interpretability by documenting decision rationale.

In the RLHF Preference Ranking Dataset project, I labeled over 5,000 response pairs for RLHF training, evaluating according to helpfulness, accuracy, safety, and coherence. My work included documenting preferences for model interpretability and maintaining high label consistency, even for complex multi-turn conversations. The project contributed to robust model evaluation and alignment with human preferences. • Performed detailed RLHF preference labeling for AI assistant evaluation • Evaluated responses with multi-criteria analysis and documentation • Managed consistency for diverse and challenging conversational data • Facilitated model interpretability by documenting decision rationale.

2023 - Present

Multilingual Text Classification Dataset Project

TextClassification
For the Multilingual Text Classification Dataset project, I created a labeled dataset of over 10,000 text samples in four languages using a multi-label classification scheme. I developed annotation guidelines, trained annotators for consistency, and achieved significant inter-annotator agreement improvements. The project supported NLP research and model evaluation by delivering high-quality multilingual data. • Designed and managed multi-label annotation workflows • Ensured label consistency with clear guidelines and training • Achieved 0.87 Cohen's Kappa score for inter-annotator agreement • Supported four languages: Polish, Russian, Ukrainian, and English.

For the Multilingual Text Classification Dataset project, I created a labeled dataset of over 10,000 text samples in four languages using a multi-label classification scheme. I developed annotation guidelines, trained annotators for consistency, and achieved significant inter-annotator agreement improvements. The project supported NLP research and model evaluation by delivering high-quality multilingual data. • Designed and managed multi-label annotation workflows • Ensured label consistency with clear guidelines and training • Achieved 0.87 Cohen's Kappa score for inter-annotator agreement • Supported four languages: Polish, Russian, Ukrainian, and English.

2023 - Present

Content Moderator & Data Reviewer

TextClassification
While working as a Content Moderator & Data Reviewer, I reviewed and classified user-generated content across multiple platforms for safety and category labels. I maintained documentation on moderation decisions, tracked ambiguous cases, and provided policy feedback to enhance guideline clarity. My efforts led to high consistency in quality audits and improved training practice documentation. • Labeled content for safety, policy violation, and theme classification • Maintained detailed records for guideline refinement and training • Achieved high consistency while processing 1000+ items weekly • Identified patterns to propose improvements in classification policies.

While working as a Content Moderator & Data Reviewer, I reviewed and classified user-generated content across multiple platforms for safety and category labels. I maintained documentation on moderation decisions, tracked ambiguous cases, and provided policy feedback to enhance guideline clarity. My efforts led to high consistency in quality audits and improved training practice documentation. • Labeled content for safety, policy violation, and theme classification • Maintained detailed records for guideline refinement and training • Achieved high consistency while processing 1000+ items weekly • Identified patterns to propose improvements in classification policies.

2023 - Present

Freelance Data Annotation & Quality Control

TextClassification
As a Freelance Data Annotation & Quality Control specialist, I labeled diverse text datasets including text classification, named entity recognition, sentiment analysis, and intent detection. I created and maintained annotation guidelines, implemented double-blind annotation, and introduced automated validation scripts. My work ensured dataset consistency, logical format compliance, and high-quality output for clients. • Processed and cleaned raw data prior to annotation using Python • Delivered labeled datasets in JSON, CSV, and XML formats • Developed and refined annotation schemes and quality checks • Measured inter-annotator agreement and improved label reliability.

As a Freelance Data Annotation & Quality Control specialist, I labeled diverse text datasets including text classification, named entity recognition, sentiment analysis, and intent detection. I created and maintained annotation guidelines, implemented double-blind annotation, and introduced automated validation scripts. My work ensured dataset consistency, logical format compliance, and high-quality output for clients. • Processed and cleaned raw data prior to annotation using Python • Delivered labeled datasets in JSON, CSV, and XML formats • Developed and refined annotation schemes and quality checks • Measured inter-annotator agreement and improved label reliability.

2023 - Present

Education

Z

ZS im. J. i W. Zamoyskich

Diploma, Information Technology

Diploma
2019 - 2024
W

WSB Merito University

Bachelor of Science, Computer Science

Bachelor of Science
2024

Work History

S

Self-Employed

VPN Solutions Engineer

Poznan
2023 - Present
S

Self-Employed

Integration & Automation Developer

Poznan
2023 - Present