Andrey Spiridonov

I am a Lead Data Scientist at a leading Customer Data Platform in hospitality with a background in Applied Mathematics and Cybernetics. I combine hands-on model development with team leadership to deliver production ML solutions that drive business impact.

I have designed and deployed demand-forecasting pipelines (DeepAR, Prophet), entity-resolution systems built on OpenSearch and approximate similarity joins, and hybrid recommendation engines. I also built RAG-based knowledge solutions using LangChain, OpenSearch and AWS Bedrock, and routinely create SageMaker pipelines, CI/CD practices, and monitoring for reliable ML products.

Earlier, I contributed to AutoML and algorithm implementations at H2O.ai and created scalable Spark/Scala pipelines for large-scale entity resolution using locality-sensitive hashing and graph clustering. My academic work includes research on evolutionary optimization for distributed ML and contributions to open-source AutoML repositories.

I mentor and hire engineers, author a Jupyter-based Applied Data Science book, and found an experimental startup to explore modern GenAI and MLOps patterns. I bring a blend of research rigor, engineering discipline, and product thinking to build sustainable, explainable, and impactful ML solutions.