For employers

Hire this AI Trainer

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

Invite to Job
Z
Zhiji L.

Zhiji L.

used to be an it engineer ,well known about computer skills

Hong Kong flagbeijing, Hong Kong

Key Skills

Software

No software listed

Top Subject Matter

soft engineer

Top Data Types

No data types listed

Top Task Types

No task types listed

Freelancer Overview

My experience primarily involves Reinforcement Learning from Human Feedback (RLHF) for LLMs. Rather than simple categorization, I evaluate model outputs based on criteria like helpfulness, harmlessness, and honesty. For instance, I often rewrite suboptimal responses or rank multiple completions to teach the model stylistic nuances. I treat each task as a teaching session: even when rejecting a response, I provide a detailed justification why. This meticulous approach to preference labeling is crucial for aligning AI behavior with complex human values

Labeling Experience

My experience focuses on ensuring high-quality training data for supervised learning models

My experience focuses on ensuring high-quality training data for supervised learning models. In my previous labeling tasks (e.g., image recognition or text classification), I strictly adhered to guideline protocols while paying close attention to edge cases—such as ambiguous lighting in photos or sarcasm in text. I learned that consistency is more important than speed; a single mislabeled data point can mislead the entire model. Therefore, I always double-check my work to maintain a high inter-rater reliability, and I actively provide feedback to requesters when the labeling rules are unclear

Not specified

Education

H

Has a bachelor's degree

Degree not specified

Not specified
Not specified

Work History

S

supervised learning models

My experience focuses on ensuring high-quality training data

Location not specified
Not specified