Joshua C.

Key Skills

Software

Other

Top Subject Matter

Model training/quantitative financial data

LLM evaluation/prompt optimization

AI output evaluation/multilingual

Top Data Types

Text

Image

Document

Top Task Types

Fine-tuning

Segmentation

Classification

Text Generation

Question Answering

Text Summarization

Data Collection

Evaluation/Rating

Freelancer Overview

I have hands-on experience with AI training data through building AI-driven automation pipelines, data-enrichment workflows, and structured database processes. As a Business Analyst, I also prepared, cleaned, and structured data for RAG workflows, while improving LLM response reliability by 90% through benchmarking, prompt optimization, and workflow testing. I also built an AI-based crypto trading system using cleaned historical market data and a self-trained LSTM model. My key strengths include attention to detail, Python/SQL, data structuring, LLM benchmarking, and an interdisciplinary background in engineering, business, and AI, which enables me to understand both technical model requirements and real business use cases.

Labeling Experience

AI-Based Crypto Quant Trading System

OtherTextSegmentation

Joshua collected, cleaned, and structured large sets of historical crypto market data for use in LSTM model training cycles. He designed, coded, and fine-tuned an LSTM-based model to generate predictive trading signals using labeled datasets. He implemented and iteratively refined the data labeling and training process using Python, Claude Code, and Codex. • Built structured datasets for AI model training and testing • Conducted full-cycle LSTM model training with labeled financial data • Automated labeling and preprocessing using Python and AI-assisted coding • Evaluated model outputs against historical market data for continuous improvement

2025 - Present

AI Business Analyst

DocumentText Generation

Joshua improved the reliability of large language model (LLM) responses by conducting benchmarking, prompt optimization, and workflow testing across internal AI tools. He evaluated model-generated outputs within real business process pipelines, focusing on response quality, relevance, and consistency. He iteratively tested, rated, and reported on LLM response data within the organization's AI-driven automation pipelines. • Designed evaluation tests and benchmarks for LLM responses • Rated and annotated model outputs for accuracy and consistency • Optimized prompts based on evaluation feedback • Supported integration of LLMs into workflow automation processes

2024 - 2026

Education

Work History