I've delivered production ML, LLM, and RAG solutions at EPAM Systems for enterprise clients, building retrieval pipelines, LLM workflows, evaluation systems, APIs, and operational tooling.
At AWS, I productionized cloud machine learning systems serving millions of requests, optimizing inference latency, throughput, reliability, and infrastructure cost. Earlier at Bitdefender, I built malware classification and anomaly-detection models from security telemetry on highly imbalanced datasets.
I build across the full AI lifecycle—from data pipelines and model training to deployment, monitoring, retraining, and production support—using Python, PyTorch, TensorFlow, LangChain, AWS, Docker, Kubernetes, and .NET.
