John Alier

Key Skills

Software

Scale AI

Mercor

Label Studio

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Question Answering

Text Summarization

RLHF

Evaluation Rating

Prompt Response Writing SFT

Segmentation

Cuboid

Object Detection

Action Recognition

Freelancer Overview

I am an experienced software engineer with a strong background in AI evaluation, data annotation, and large-scale training data pipelines. My work includes daily preference ranking and comparative judgment of multimodal AI outputs—spanning images, audio, and video—to ensure model alignment and quality. I am highly skilled in using specialized annotation platforms and RLHF (Reinforcement Learning with Human Feedback) techniques to support the fine-tuning of generative models. I thrive in cross-functional teams, collaborating closely with AI researchers to identify trends, provide actionable feedback, and maintain high accuracy and throughput in annotation workflows. My technical expertise covers Python, SQL, cloud platforms (AWS, GCP, Azure), and advanced data modeling, allowing me to bridge the gap between engineering and precise, reliable AI training data.

Labeling Experience

Text, Image

MercorTextRLHF

Evaluated AI-generated content for quality, safety, policy adherence, and usability. Labeled outputs based on predefined rubrics covering correctness, tone, relevance, and potential risk. Flagged unsafe, misleading, or low-quality outputs and provided structured feedback to improve downstream model behavior. Ensured consistency and reliability across large evaluation batches.

Evaluated AI-generated content for quality, safety, policy adherence, and usability. Labeled outputs based on predefined rubrics covering correctness, tone, relevance, and potential risk. Flagged unsafe, misleading, or low-quality outputs and provided structured feedback to improve downstream model behavior. Ensured consistency and reliability across large evaluation batches.

2025

AI Evaluation & Image Data Labeling (RLHF)

Label StudioImageRLHF

Performed RLHF and multimodal evaluation to support the alignment and reliability of next-generation AI models. Carefully reviewed image-based question-and-answer items, verifying and correcting labels according to predefined guidelines. Applied comparative judgment and preference ranking techniques for diverse multimodal model outputs. • Maintained high throughput and accuracy while working remotely on detailed guideline-driven tasks. • Extracted exact text from images to support consistent QA and labeling decisions across a variety of visual content. • Flagged unanswerable or policy-infringing items and documented clear, concise rationale for each. • Collaborated with research teams to provide actionable feedback and identify key quality trends.

Performed RLHF and multimodal evaluation to support the alignment and reliability of next-generation AI models. Carefully reviewed image-based question-and-answer items, verifying and correcting labels according to predefined guidelines. Applied comparative judgment and preference ranking techniques for diverse multimodal model outputs. • Maintained high throughput and accuracy while working remotely on detailed guideline-driven tasks. • Extracted exact text from images to support consistent QA and labeling decisions across a variety of visual content. • Flagged unanswerable or policy-infringing items and documented clear, concise rationale for each. • Collaborated with research teams to provide actionable feedback and identify key quality trends.

2021

Image data labeling

LabelboxImageSegmentationCuboid

Performed image data labeling and QA for computer vision datasets using Labelbox, focusing on high-precision annotation and consistent label application across large batches. Created and reviewed annotations for object detection tasks using bounding boxes, polygons, and keypoints, ensuring tight alignment with project guidelines and visual evidence. Verified and corrected labels, extracted exact text from images when required, and produced clear, standalone notes for edge cases or unanswerable items. Maintained strong attention to detail, grammar, and formatting in all outputs while meeting productivity targets and collaborating through iterative feedback to improve annotation quality.

Performed image data labeling and QA for computer vision datasets using Labelbox, focusing on high-precision annotation and consistent label application across large batches. Created and reviewed annotations for object detection tasks using bounding boxes, polygons, and keypoints, ensuring tight alignment with project guidelines and visual evidence. Verified and corrected labels, extracted exact text from images when required, and produced clear, standalone notes for edge cases or unanswerable items. Maintained strong attention to detail, grammar, and formatting in all outputs while meeting productivity targets and collaborating through iterative feedback to improve annotation quality.

2022 - 2024

Multimodal RLHF Evaluation for Generative AI Models

Scale AITextQuestion AnsweringText Summarization

Conducted large-scale RLHF-based evaluation of generative AI outputs across text, image, audio, and video modalities. Performed detailed preference ranking and comparative judgments against prompts to assess relevance, coherence, factuality, and alignment with human intent. Maintained high accuracy and throughput while adhering to strict annotation guidelines and quality standards. Collaborated with AI research teams to surface systematic quality trends and provide actionable feedback used in model fine-tuning and alignment.

Conducted large-scale RLHF-based evaluation of generative AI outputs across text, image, audio, and video modalities. Performed detailed preference ranking and comparative judgments against prompts to assess relevance, coherence, factuality, and alignment with human intent. Maintained high accuracy and throughput while adhering to strict annotation guidelines and quality standards. Collaborated with AI research teams to surface systematic quality trends and provide actionable feedback used in model fine-tuning and alignment.

2021 - 2024

Education

Work History