AI Trainer and Data Labeler (Mindrift / Toloka / Multiple Platforms)
MindriftTextRLHF
Oversaw AI training and data annotation tasks across multiple platforms, focusing on large language model (LLM) validation and structured feedback. Evaluated LLM-generated outputs for factual correctness, reasoning, and logical consistency to improve model quality and reliability. Delivered systematic annotation for RLHF, preference ranking, and SFT labeling while enforcing strict data quality standards. • Evaluated and scored LLM outputs in STEM, data science, and ESG domains
• Validated the PSX ESG Controversies dataset using a Python and LLM-assisted pipeline
• Utilized tools such as Labelbox, CVAT, Label Studio, and proprietary annotation platforms
• Maintained reproducible outputs, consistency checks, and documentation
Oversaw AI training and data annotation tasks across multiple platforms, focusing on large language model (LLM) validation and structured feedback. Evaluated LLM-generated outputs for factual correctness, reasoning, and logical consistency to improve model quality and reliability. Delivered systematic annotation for RLHF, preference ranking, and SFT labeling while enforcing strict data quality standards. • Evaluated and scored LLM outputs in STEM, data science, and ESG domains • Validated the PSX ESG Controversies dataset using a Python and LLM-assisted pipeline • Utilized tools such as Labelbox, CVAT, Label Studio, and proprietary annotation platforms • Maintained reproducible outputs, consistency checks, and documentation
2023 - Present