Caleb K.

Key Skills

Software

Label Studio

Top Subject Matter

Multilingual AI content evaluation and annotation

Multilingual image annotation for vision models and autonomous systems

Computational linguistics research annotation and ML evaluation for low-resource languages

Top Data Types

Text

Image

Top Task Types

Bounding Box

Segmentation

Classification

Red Teaming

Entity (NER) Classification

Freelancer Overview

AI Language Data Specialist (Multilingual Focus), Freelance Independent Contractor (Remote). Brings 9+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Hugging Face datasets. Education includes Doctor of Philosophy, Harvard University (2023) and Master of Science, Harvard University (2018). AI-training focus includes data types such as Text and Image and labeling workflows including Evaluation, Rating, and Bounding Box.

Labeling Experience

Multilingual AI Evaluation Specialist (Freelance) - Independent Contractor

Label StudioImageEntity (NER) Classification

Freelance work focused on evaluating and benchmarking multilingual AI outputs using structured comparison and quality assurance methods. Responsible for planning evaluation workflows, maintaining label consistency, and delivering reliable results to internal or partner stakeholders. Required strong attention to detail, multilingual expertise, and the ability to operate independently under clear annotation and evaluation protocols. • Completed multilingual pairwise comparison tasks to assess AI outputs against human reference standards. • Produced large-scale sentence-level annotations supporting semantic similarity and grammatical acceptability evaluation. • Conducted audio and multimedia preprocessing workflows including transcription and noise labeling support. • Performed image/text quality assurance audits and collaborated with distributed teams to meet weekly throughput targets.

2023 - Present

AI Language Data Specialist (Multilingual Focus), Freelance Independent Contractor (Remote)

Text

Produced multilingual pairwise comparisons by comparing AI-generated outputs to human reference standards across five languages and maintaining 98% consistency with gold labels. Labeled text for semantic similarity and grammatical acceptability to support internal language model evaluations. • Pairwise comparisons of model outputs vs gold references • Semantic similarity and grammatical acceptability labeling • Linguistic counting/frequency tasks to support grammar rule extraction • Bias-focused evaluation rubric application for safety and relevance

2023 - Present

Multilingual AI Safety Evaluation Simulation (Independent Project)

TextRed Teaming

Ran a multilingual AI safety evaluation simulation that generated model responses from prompts and then tagged outputs as unsafe or irrelevant. Documented reusable annotation guidelines covering edge cases to support consistent safety labeling. • Safety labeling (unsafe vs irrelevant) for multilingual outputs • Prompt-to-response evaluation via pairwise comparisons • Inter-annotator agreement tracking (0.89) • Authored a multi-lingual content safety labeling protocol

2024 - 2024

Computational Linguistics Researcher (PhD) - Harvard University

Label StudioTextSegmentation

PhD research focused on language technologies and low-resource language evaluation methods, including the design and testing of evaluation and annotation schemes. Responsible for building gold-standard datasets, running rigorous experiments to evaluate model behavior across multiple languages, and automating quality checks to support large-scale analysis. Required advanced research skills in computational linguistics, experimental design, and robust programming for data validation. • Designed annotation schemes for morphological segmentation and produced tokenized datasets used in academic work. • Evaluated transformer-based LLM outputs for multiple languages and developed bias detection rubrics adopted in subsequent research. • Led pairwise fluency evaluation studies on machine-translated text and recommended post-editing strategies. • Automated data quality checks for large labeled sets using Python validation and regex-based scripts to reduce manual review time.

2018 - 2023

Computational Linguistics Researcher, Harvard University (PhD; Language Technologies Lab)

TextSegmentation

Created morphological segmentation annotation schemes and produced gold-standard datasets with tokenized words across multiple languages. Led evaluation work for transformer-based LLM outputs by developing bias-detection rubrics and running large-scale automated quality checks. • Morphological segmentation schema creation and tokenized dataset labeling • Bias detection rubric development for evaluated outputs • Pairwise comparison studies for machine translation fluency errors • Automated QA for 50,000+ labeled instances using validation scripts

2018 - 2023

Show more

Education

Work History