Kelvin N.

Key Skills

Software

Remotasks

Top Subject Matter

LLM prompt engineering

AI output evaluation

training data QA

Top Data Types

Text

Image

Document

Top Task Types

Prompt + Response Writing (SFT)

Text Generation

Transcription

Data Collection

Classification

Freelancer Overview

AI Trainer & Evaluator (Freelance / Self-Directed). Brings professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include : AI / LLM Tools ChatGPT / GPT-4, Claude (Anthropic), Gemini, Midjourney, DALL-E Writing & Editing Technical writing, fact-checking, style-guide adherence, C1/C2 English proficiency Evaluation AI output rating, annotation, RLHF-style feedback, logical reasoning review, QA Data & Tech Data workflows, structured content tasks, spreadsheet analysis, version control concepts Productivity Notion, Google Workspace, Markdown, Microsoft Office In Progress SQL for data analysis (beginner) — actively building proficiency through hands-on practice Education includes AI Image Generation — AI-training focus includes data types such as Text, Image, and Document and labeling workflows including Evaluation, Rating, and Prompt + Response Writing (SFT).

Labeling Experience

AI-Assisted Technical Documentation (Freelance Project)

OtherDocumentText Generation

Used LLMs to draft, review, and improve technical documentation with structured editing for logical flow. Applied fact-checking and style-guide adherence to ensure clarity and factual accuracy in deliverables. Created reusable templates and checklists to maintain consistency across multi-document projects. • Generated and revised technical guides using Claude and GPT-4 • Ensured coherence, clarity, and instruction-aligned outputs • Built templates/checklists for consistent documentation QA • Produced polished written artifacts for cross-functional stakeholders

2022 - Present

Prompt Engineering & Model Evaluation (Self-Directed Project)

OtherTextRed Teaming

Tested hundreds of prompt variations across multiple LLMs to probe instruction-following, reasoning, and creative generation behavior. Evaluated model outputs with custom rubrics across dimensions including accuracy, fluency, relevance, and safety. Maintained a personal knowledge base of prompting strategies for quality improvements. • Used rubric-based evaluations for RLHF applicability • Performed safety-oriented red-teaming style assessment • Compared outcomes across GPT-4, Claude, and Gemini • Documented prompt patterns for reliable behaviors

2022 - Present

AI Trainer & Evaluator (Image generation prompting)

OtherImagePrompt + Response Writing (SFT)

Generated and critiqued AI image outputs by creating detailed instruction-following prompts for image generation models. Iterated on prompt wording based on output quality assessments to improve visual instruction fidelity. Demonstrated precision describing cinematic and creative visual requirements. • Wrote text-to-image prompts for DALL-E and Midjourney • Evaluated image results for instruction adherence • Refined prompt parameters through multiple iteration cycles • Translated abstract concepts into structured visual descriptions

2021 - Present

AI Trainer & Evaluator (Freelance / Self-Directed)

Don't discloseText

Evaluated AI-generated text for accuracy, coherence, logical consistency, and alignment with task instructions using quality rubrics. Produced structured, human-readable training samples spanning multiple domains including technical writing, creative content, and factual Q&A. Iteratively reviewed and refined model outputs by providing actionable written feedback to improve response quality. • Rated outputs across accuracy, fluency, relevance, and nuance • Applied RLHF-style evaluation and annotation workflows • Assessed instruction-following and logical reasoning quality • Crafted prompt-response examples for training data

2021 - Present

Education

Work History