I am an AI trainer and evaluator with 25 years of practical experience across enterprise IT, SAP ERP, logistics, supply chain, procurement, and Lean process improvement. I combine domain expertise with hands-on AI data training, response evaluation, annotation, research, and fact-checking.
I work on video and audio annotation, rubric scoring, preference ranking, web research, transcription, prompt writing, and guideline adherence for AI data platforms. My background enables me to evaluate whether model outputs are operationally accurate as well as clear and useful.
I have extensive supply-chain experience in automotive and food-processing environments, including logistics planning, stock control, SAP rollouts, Lean implementation, PDCA, 5S, and process optimization. I have managed workflows from raw-material intake through finished-goods dispatch.
I also build agentic AI systems. I created aura-code, an open-source TypeScript multi-agent framework with LLM streaming, tool execution, human-in-the-loop confirmation, permission controls, safety mechanisms, CI/CD, and extensive automated testing.
Alongside my technical and operational career, I have more than eleven years of B2B English teaching experience. I design curricula and assessments, teach business communication, and provide detailed written feedback for corporate professionals.
I am a native Serbian speaker with fluent English proficiency and working knowledge of German and Spanish. I bring a rigorous, multilingual, systems-oriented perspective to AI training, enterprise workflows, and applied model evaluation.