Vinoth V

• Led end-to-end program delivery for AWS AI/ML flagship services including Amazon Rekognition, MTurk, Bedrock, and SageMaker, managing cross-functional teams and $10M+ annual budgets with 99.9% SLA compliance.

• Scaled global operations from 50 to 500+ resources through strategic vendor partnerships, implementing standardized quality frameworks that improved productivity by 25% and reduced operational costs by 30%.

• Developed and implemented career frameworks for 200+ employees across multiple locations, achieving 90% retention rate while successfully transitioning 300+ team members to new workflows with zero business disruption.

• Established automated quality assurance systems for ML operations, resulting in 50% reduction in annotation errors, 70% improvement in model training efficiency, and 40% faster project delivery times.

• Drove strategic OP1-2 initiatives and service launches, developing data-driven frameworks that enhanced visibility across 10+ concurrent ML projects.

• Managed large-scale program delivery for AWS AI/ML language and document services, overseeing complex capacity forecasting and vendor strategies with multimillion-dollar P&L responsibility.

• Implemented innovative annotation workflows and security protocols, achieving 80% throughput improvement through enhanced tooling and engineering collaboration.

• Established technical frameworks and SOPs incorporating scientist feedback loops, improving model accuracy and cross-functional collaboration.

• Directed quality assurance initiatives for ML data workflows, executing audits and instituting process enhancements.
• Developed scalable SOPs for annotation workflows across Echo Look, Cloud Cam, and Ring devices.
• Updated ML workflow documentation, contributing to improved team efficiency. Realised targeted KPIs for productivity, quality, and utilisation across various projects.

• Executed ground truth data collection initiatives for product development, ensuring 100% accuracy in documentation.
• Implemented testing procedures to uphold quality benchmarks across repetitive tasks.