For employers

Hire this AI Trainer

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

Invite to Job
Kyle Hsiao

Kyle Hsiao

LLM Eval and Text Gen Specialist in Traditional Chinese (Taiwan)

Taiwan flagNew Taipei City, Taiwan
$30.00/hrIntermediateAppenData Annotation TechOneforma

Key Skills

Software

AppenAppen
Data Annotation TechData Annotation Tech
OneFormaOneForma
RemotasksRemotasks
Scale AIScale AI
Other

Top Subject Matter

No subject matter listed

Top Data Types

DocumentDocument
ImageImage
TextText

Top Task Types

Evaluation Rating
Prompt Response Writing SFT
Question Answering
Red Teaming
Translation Localization

Freelancer Overview

An expert in LLM text analysis. I was part of safety risk, localization, SFT, RLHF projects for various clients and was promoted to manager for my great quality since I'm detail-oriented and sticks to the instructions really well. In addition, given my translation/localization background, I show deep understanding of my language, Traditional Chinese (Taiwan), which helps a lot when dealing with LLL generated text that is not usually properly localized.

IntermediateEnglishChinese Mandarin

Labeling Experience

Data Annotation Tech

Data Annotator

Data Annotation TechTextTranslation LocalizationEvaluation Rating
Mostly evaluating LLM generated text according to different factors and rewrite/localize them if necessary.

Mostly evaluating LLM generated text according to different factors and rewrite/localize them if necessary.

2024
Scale AI

Contributor/Manager

Scale AITextTranslation LocalizationRLHF
There are various projects that I worked with Scale AI. All of them has corresponding instruction documents to follow. Due to NDA, I will not be able to share too much detail. * Translation/Localization - We check if the machine-translated texts are good. If not, we localized them properly. * RLHF - We improve the LLM responses with text that is better all rounds iteratively. * Evaluation/Rating - We rate 2 LLM responses according to several dimensions and choose a better one. * Red Teaming - We identify and fix the possible risk in the LLM response in regards to sensitive subjects. * SFT - We create a ideal prompt and response pair and compare it to the LLM response.

There are various projects that I worked with Scale AI. All of them has corresponding instruction documents to follow. Due to NDA, I will not be able to share too much detail. * Translation/Localization - We check if the machine-translated texts are good. If not, we localized them properly. * RLHF - We improve the LLM responses with text that is better all rounds iteratively. * Evaluation/Rating - We rate 2 LLM responses according to several dimensions and choose a better one. * Red Teaming - We identify and fix the possible risk in the LLM response in regards to sensitive subjects. * SFT - We create a ideal prompt and response pair and compare it to the LLM response.

2024

Education

N

National Tsing Hua University

Bachelor of Science, Physics, Computer Science

Bachelor of Science
2018 - 2023

Work History

S

SuccessGlo PTE Ltd.

Freelance Translator

Singapore
2022 - Present
L

Logitech FE Ltd.

CMF Engineering Intern

Hsinchu
2022 - 2022