Tech Lead for Unistore (Hybrid tables), which is Snowflake's unified solution for analytical and transactional processing (HTAP). My contributions encompass, but are not limited to, Hybrid Tables' plan compilation, execution engine, access path selection, storage medium selection, constraint enforcement, consistency and concurrency control algorithms/protocols, index support, online DDLs, SQL features, and query processing. As a founding team member, I have been instrumental in, and contributed to, the majority of the architectural decisions.

Hybrid Tables offer exceptional price-performance for both transactional and analytical needs, enabling thousands of customers to accommodate their agentic AI workloads.

Nikos was a member of the AWS Redshift's Query Processing team. His work focused on re-architecting Redshift's components to improve availability, performance, and fault tolerance. His work focused on metadata management, data-sharing, and time-travel for metadata.

Development and deployment of tools for large-scale data analysis in distributed environments. In detail, Nikos' work focused on the adaptability of distributed Stream Processing Engines (i.e., Apache Flink, Spark Streaming, Apache Storm) and the trade-off between performance and accuracy, load balance, and scaling-out/-in operations. On top of this, Nikos deployed visualization tools for big graphs in Causal Discovery (https://www.ccd.pitt.edu/). Nikos' work was conducted in the context of public-Cloud deployments, such as AWS and Microsoft Azure.

Part-time Instructor for CS1555 Database Management Systems course. Nikos' responsibilities were: creating and grading assignments, conduct recitation lectures, term project creation/grading.

Nikos worked on the Streams project of the Database Group of Microsoft Research, in Redmond WA. In detail, Nikos extended Microsoft's Trill streaming engine to dynamically generate code for efficient parsing of semi-structured data. On top of this, Nikos developed a parsing algorithm for semi-structured data, Mison, which improves JSON parsing speeds up to 10x compared with the state-of-the-art solutions.