Andile D.

Key Skills

Freelancer Overview

I have not yet worked in a formal AI training-data or data-labeling role, but I have practical experience evaluating and validating AI outputs through applied AI engineering projects. In my Natural Language Interface to Database project, I built an LLM evaluation harness for Text-to-SQL, comparing generated SQL against expected outputs using metrics such as Exact Match, Levenshtein distance, BLEU and ROUGE. This gave me experience assessing model responses for correctness, consistency and usefulness. I also built Pulse, an AI-assisted circuit design and simulation agent that converts natural-language circuit requests into structured schemas, SPICE netlists, ngspice simulations, plots and evaluation feedback. This project involved prompt design, structured outputs, schema validation, tool use and feedback loops. These experiences have strengthened my ability to judge AI-generated outputs, identify errors, apply evaluation criteria and improve the reliability of AI workflows.

Education

Work History