At Tabiya, I architect GenAI systems with LangGraph and Azure OpenAI Service, automating lead scoring and CRM logic to reduce manual operational tasks by 37%.
I also designed a BigQuery data warehouse and event-driven processing on GCP, achieving over 90% match accuracy across 17 behavioral domains. I optimized Azure RAG and MLOps architectures to support 99.9% reliability for production LLM deployments.
Earlier at Tabiya, I enhanced agentic LLM workflows and localized NLP solutions, including a Portuguese chatbot that reduced international support volume by 68%. I also worked on an ICP analytics platform that achieved an 86% match rate for real-time lead scoring and customer segmentation.
At StyleSeat, I scaled production ML infrastructure and improved data processing for marketplace analytics, increasing Redshift indexing throughput 16x. At Regeneron, I improved clinical speech-to-text accuracy by 52% and automated clinical document extraction.
