AI Data Service Safety Model Operations Team Lead
TikTok
London, England
07.2024 - Current
- Directed and delivered 8 AI safety model operations projects simultaneously (LLM, AI Moderation), consistently achieving >95–99% accuracy across global markets.
- Supported safe deployment of Large Language Models (LLMs) by iterating on evaluation frameworks, refining calibration methods, and embedding machine learning safety checks into QA processes.
- Launched, successfully led and brought to completion and on target AI Moderation project, delivering 95.59% accuracy from inception while managing cross-project dependencies and resourcing constraints.
- Improved underperforming markets in LLM Photo specific machine learning project through Get 2 Green planning, boosting AI model precision by +17% (FR) and +6% (IT) within weeks.
- Partnered with Subject Matter Experts, In house Trainers, and Project Managers to enhance AI model iteration, feeding QA insights and RCA data into continuous model refinement.
- Built and led a high-performing Analyst team, with Subject Matter Experts and in house Trainer positions (20+ members) with sustained >99% accuracy, improving model evaluation quality and reliability.
- Strengthened global stakeholder collaboration (POCs, POMs, SMEs, BPOs), ensuring smooth alignment and early escalation of systemic issues in AI projects.
- Enhanced data analysis and trend monitoring processes to optimize QA performance, reduce errors, and reinforce AI safety alignment at scale.
- Ramp up of multiple markets across 8 projects at multiple global Business Partner Outsourcing sites.
- Supported transition of AI Data team into AI Model Operations team specialising in platform safety - Managing 43 QAs, 1 SME and 1 Trainer
