At Value Driven Data, I led delivery of an AI-powered demand-planning and waste-management system with a six-engineer team. The system cut store-level wastage 17% and lifted gross margin 3%.
I defined architecture, technical strategy, and delivery plans across client engagements. I also presented in recurring business and technical reviews with client executives, store managers, and operations managers.
For a crypto investment advisory startup, I built blockchain data-ingestion pipelines for smart-contract intelligence and designed a RAG-based investment advisor chatbot using live market signals. For a national postal and logistics operator, I built a natural-language analytics chatbot and supervised routing optimization algorithms.
Previously, at Omaline, I built an LLM-powered semantic search engine and ranking algorithms, and fine-tuned language models on GPU infrastructure. I also contributed the ConjugateTranspose operation to the TensorFlow frontend of Intel OpenVINO Toolkit.
