Ajay Krishna

WORK HISTORY

EDUCATION

ABOUT AJAY KRISHNA

I turn unstructured, high-volume data into systems that don\'t just process — they reason, reconcile, and explain. That\'s what I\'ve done across 3+ years of enterprise data leadership for Fortune 500 clients. That\'s what I\'m now building end-to-end as an AI and GenAI Engineer. At Gameopedia, I led data quality operations across pipelines processing 5M+ daily records for Samsung, Google, Xbox, and Unity. Managed a team of 50+, defined validation frameworks, and served as the technical bridge between client stakeholders and engineering teams. The work was invisible when it was working — which, by design, was always. After moving to Canada, I completed postgraduate certifications in AI and Data Analytics at Durham College — then kept building. A cloud-native forecasting platform on AWS with Airflow orchestration, Snowflake warehousing, Medallion architecture, and XGBoost predictions served via FastAPI at sub-200ms latency. A real-time Kafka streaming pipeline with drift detection and automated model retraining. And most recently, PolicyIntel — a multi-document RAG engine on Azure that reconciles conflicting policy documents, flags risk levels, and returns every answer with a page-level citation and a full audit trail. Built it. Shipped it. Moving on to the next one. The shift from data engineering to GenAI wasn\'t a restart. The years of pipeline work are exactly why I understand why chunk size matters, why deduplication breaks retrieval if you skip it, and why an audit trail isn\'t optional in a regulated industry. It\'s the same instinct — reliability, traceability, trust in the output — just applied to a system that thinks. Core Stack: Python · Azure OpenAI · Azure AI Search · Azure Document Intelligence · RAG · LangChain · Multi-Agent Systems · FastAPI · Apache Kafka · Airflow · Snowflake · AWS · Docker · PostgreSQL · Streamlit Open to GenAI Engineer · AI Engineer · LLM Engineer · Applied AI Engineer roles in Ontario.