AI/ML Analyst (Innodata for Meta)
TextRLHF
As an AI/ML Analyst at Innodata (Client: Meta), I evaluated over 500 LLM responses per day using RLHF-based evaluation. My responsibilities included ranking, comparison, and justification writing to ensure the quality of model responses. Hallucination detection and strict SOP-driven QA formed a critical aspect of my role.
• Evaluated and rated prompts and responses for accuracy and compliance.
• Identified hallucinations, edge cases, and flagged policy violations.
• Maintained annotation quality through process adherence.
• Utilized proprietary/internal evaluation platforms and tools.
As an AI/ML Analyst at Innodata (Client: Meta), I evaluated over 500 LLM responses per day using RLHF-based evaluation. My responsibilities included ranking, comparison, and justification writing to ensure the quality of model responses. Hallucination detection and strict SOP-driven QA formed a critical aspect of my role. • Evaluated and rated prompts and responses for accuracy and compliance. • Identified hallucinations, edge cases, and flagged policy violations. • Maintained annotation quality through process adherence. • Utilized proprietary/internal evaluation platforms and tools.
2025 - Present