Ekta Sojitra
Ekta Sojitra
Data Analytics and Data Science Engineer II, PG&E
Title of Talk:
First Foundations for Trustworthy AI on Industrial Time-Series Data
Abstract
Computational intelligence models trained on industrial time-series data inherit silent failures that healthy pipelines never reveal. This talk traces how temporal distortion, non-deterministic recomputation, and semantic erasure corrupt analytical foundations, and argues that reliability-first design, deterministic processing boundaries and semantic preservation is a precondition for trustworthy industrial AI.
Brief Profile
Ekta Sojitra is a Data Analytics and Data Science Engineer specializing in industrial OT/IT data integration. Based in Peoria, Illinois, she works with AVEVA PI, Oracle Data Integrator, Snowflake, and Power BI in mission-critical utility environments. Her published research spans historian-to-cloud integration, pipeline observability, and reliability-first analytics design, forming a coherent body of work on industrial time-series data correctness.