Hi, my name is

Ausiankin Kiryl.
I work with AI data.

Labeling, annotating, and preparing data pipelines for training AI models. I work with text, images, and multimodal data.

Working on mindrift.ai

View projects Get in touch

Years of experience

1000+

Hours of annotation

50K+

Labeled items

Languages (RU / EN)

Work experience

2023 — Present

AI Data Annotator

mindrift.ai

Annotation of text and multimodal data, model response evaluation (RLHF), prompt work, and guideline compliance for training datasets.

2020 — 2023

Data Annotator

TOLOKA AI

Data annotation and labeling for AI models, working with text and visual datasets, ensuring labeling quality and consistency.

Tools & platforms

Mindrift.ai

TOLOKA AI

Label Studio

CVAT

Google Sheets

Notion

ChatGPT

Prodigy

About me

I'm an AI data specialist. I annotate and label data streams used to train and fine-tune artificial intelligence models.

I work with text, images, and multimodal data, ensuring accuracy, consistency, and compliance with project requirements. I understand how labeling quality affects model behavior, so I pay close attention to detail and follow guidelines strictly.

My work sits at the intersection of data and artificial intelligence: I help build efficient data-processing workflows so that every accurately labeled example brings a model one step closer to more precise and reliable results.

Data Annotation Data Labeling RLHF Text Annotation Image Labeling Multimodal Data QA & Consistency Prompt Evaluation

Projects

Text Data Annotation

Classification, NER, and QA of text streams for training datasets.

TextNERQA

Image Annotation

Bounding boxes, segmentation, and multimodal labeling for vision models.

VisionBBoxSegmentation

RLHF & Response Evaluation

Evaluating and ranking model responses, working with prompts and feedback.

RLHFPromptsFeedback

Data Pipelines

Building and optimizing AI data-processing workflows, following guidelines.

PipelinesGuidelinesWorkflow

Work principles

Accuracy

I strictly follow project guidelines. Every labeled example is checked against requirements to avoid errors in training data.

Consistency

I maintain a unified approach across all annotations. This is crucial for AI models — data consistency directly affects training quality.

Speed

I deliver tasks on time without compromising quality. I can handle large data volumes while staying focused and productive.

Feedback

Open to corrections and clarifications. When instructions are ambiguous, I ask the team right away instead of guessing.

Contact me

Open to new projects and opportunities. Drop me a line — I'll reply within a day.

Email 5415111@gmail.com

Phone +375 29 247-74-47

Address Khimshiashvili st.15 ap.45, Batumi, 6010, Georgia