Antony Mwenda

Key Skills

Software

Appen

Clickworker

Data Annotation Tech

Google Cloud Vertex AI

Labelbox

Mindrift

OneForma

Remotasks

Sama

Scale AI

Toloka

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Geospatial Tiled Imagery

Text

Top Task Types

Computer Programming Coding

Data Collection

Prompt Response Writing SFT

RLHF

Translation Localization

Freelancer Overview

Experienced in AI data labeling and training, I specialize in text-based tasks (NER, classification, SFT prompts) and multimodal annotation (bounding boxes, segmentation) with a focus on healthcare, geospatial, and multilingual NLP projects. My technical proficiency in Python/SQL and expertise in LLM fine-tuning (RLHF, evaluation) enable me to bridge domain knowledge with scalable labeling workflows. I stand out through precision in complex labeling (e.g., medical entity recognition, non-English prompt refinement) and experience with QA pipelines to ensure dataset integrity. Passionate about optimizing annotation guidelines and mentoring teams, I’ve contributed to high-impact AI models in regulated industries.

Labeling Experience

LLM Fine-Tuning for Scientific Text Classification

Scale AITextRLHF

From January to December 2023, I led a data labeling project focused on fine-tuning LLMs for scientific and mathematical content. The scope included: Text Classification: Categorizing research papers by discipline (e.g., biomathematics, physics) and complexity level. NER Annotation: Identifying and labeling key scientific entities (equations, theorems, datasets) in academic texts. SFT Prompt Engineering: Crafting and rating prompt-response pairs to improve LLM outputs for technical queries. QA Evaluation: Assessing model responses for accuracy, logical consistency, and relevance to STEM domains. Project Size: Annotated 5,000+ documents with a 98% inter-annotator agreement score, adhering to strict guidelines for technical precision. Implemented iterative QA checks to resolve edge cases (e.g., ambiguous mathematical notation).

From January to December 2023, I led a data labeling project focused on fine-tuning LLMs for scientific and mathematical content. The scope included: Text Classification: Categorizing research papers by discipline (e.g., biomathematics, physics) and complexity level. NER Annotation: Identifying and labeling key scientific entities (equations, theorems, datasets) in academic texts. SFT Prompt Engineering: Crafting and rating prompt-response pairs to improve LLM outputs for technical queries. QA Evaluation: Assessing model responses for accuracy, logical consistency, and relevance to STEM domains. Project Size: Annotated 5,000+ documents with a 98% inter-annotator agreement score, adhering to strict guidelines for technical precision. Implemented iterative QA checks to resolve edge cases (e.g., ambiguous mathematical notation).

2023 - 2023

Education

Work History