Mark W.

Key Skills

Software

No software listed

Top Subject Matter

AI

HealthCare

Top Data Types

Text

Document

Image

Top Task Types

Entity (NER) Classification

Classification

Text Generation

Question Answering

RLHF

Fine-tuning

Text Summarization

Data Collection

Freelancer Overview

I specialize in the architecture and optimization of high-quality training data ecosystems for enterprise-grade AI. My expertise lies in designing the data foundations for LLM and RAG systems, where I have led the end-to-end pipeline from raw document processing to structured knowledge retrieval. At Azeus Systems, I engineered the data ingestion and labeling logic for an AI4RFP automation platform, transforming unstructured corporate documents into clean, vectorized datasets. This involved developing OCR extraction workflows and data cleaning protocols that ensured the training data was accurate and contextually relevant, directly contributing to a 60% reduction in manual labor. My background in Data Science and published research on foundation models (NeurIPS 2024) provides a deep understanding of how data quality impacts model performance. Unlike traditional annotators, I bridge the gap between raw data labeling and model architecture. I have designed data schemas for specialized domains—ranging from financial trading instructions to gut microbiome analysis—ensuring that datasets are not only labeled correctly but are also optimized for scalability and inference efficiency. Proficient in Python, LangChain, and Vector Databases, I bring a strategic, engineering-first approach to building robust training datasets that drive production AI success.

Labeling Experience

Researcher

TextRLHF

Need to Finetune a LLAMA model to do classification tasks

2024 - 2025

Education

Work History