AI Trainer & Data Analytics Consultant | Freelance & Remote Contracts (2024 – Present)
Text
Fine-tuningFreelance AI Trainer & Data Analytics Consultant focused on building end-to-end training pipelines to improve domain-specific model accuracy. Conducted supervised fine-tuning–style dataset preparation by parsing unstructured inputs into structured training corpora and using few-shot + Chain-of-Thought prompting to reduce hallucinations. Applied evaluation and feedback loops to grade outputs, red-team vulnerabilities, and refine prompts for legal, operational, and conversational modules.• Built training corpora by transforming unstructured data arrays into high-fidelity datasets in OpenTrain AI.• Authored few-shot and Chain-of-Thought (CoT) prompt frameworks to improve output accuracy and reduce hallucinations by 35%.• Implemented RLHF/RLAIF-based validation and grading to support label refinement and model safety checks.• Performed red teaming and adversarial prompting to identify and mitigate model vulnerabilities.