Kalkidan Abey

Key Skills

Software

Mercor

Top Subject Matter

General/Multiple Domains

Top Data Types

Text

Image

Video

Top Task Types

RLHF

Bounding Box

Classification

Object Detection

Text Generation

Question Answering

Text Summarization

Fine Tuning

Transcription

Evaluation Rating

Prompt Response Writing SFT

Data Collection

Freelancer Overview

I have hands-on experience across multiple AI training platforms including Handshake AI, Outlier AI, and Mercor, where I contributed to a wide range of data labeling and model training tasks. My work has spanned quality control, response evaluation, model comparisons, critique and feedback writing, audio and language training, video and image generation evaluation, and landscape output assessment. Across these roles I developed the ability to assess AI outputs quickly and accurately, identify errors, bias, and inconsistencies, and deliver structured feedback that directly supports model improvement. What sets me apart is the combination of breadth and domain knowledge I bring to this work. Having contributed to generalist evaluations as well as specialized audio, language, and visual modalities, I am comfortable working across different types of AI systems and training pipelines. My background in neuroscience and ongoing research experience at Oregon Health and Sciences University also gives me a strong foundation for evaluating AI outputs in scientific and medical contexts, where accuracy and responsible communication are critical. I am detail-oriented, reliable, and genuinely invested in helping build AI systems that are safe, helpful, and high quality.

Labeling Experience

AI Evaluator & Trainer

MercorTextRLHF

Worked as an AI Evaluator & Trainer focused on improving chatbot systems. Participated in structured conversations to generate high-quality training data and evaluated AI outputs for accuracy and safety. Provided feedback to enhance model performance by identifying issues and documenting findings. • Evaluated chatbot answers for helpfulness, correctness, and user intent alignment. • Conducted qualitative analysis to spot model biases and errors. • Developed creative prompts for adversarial testing of AI systems. • Used structured forms to submit findings and enhance RLHF pipelines.

Worked as an AI Evaluator & Trainer focused on improving chatbot systems. Participated in structured conversations to generate high-quality training data and evaluated AI outputs for accuracy and safety. Provided feedback to enhance model performance by identifying issues and documenting findings. • Evaluated chatbot answers for helpfulness, correctness, and user intent alignment. • Conducted qualitative analysis to spot model biases and errors. • Developed creative prompts for adversarial testing of AI systems. • Used structured forms to submit findings and enhance RLHF pipelines.

2024 - Present

Education

Work History