Industrial optimization on Kafka, InfluxDB and AI/ML
How I split the time-series write path at Andritz with Kafka and InfluxDB for 3x faster queries, and shipped AI/ML optimization that cut manual calibration time 40%.
Constraint
Time-series reads from plant sensors were degrading in SQL Server, and calibration still depended on manual operator adjustment.
Decision
Split the write path: Kafka streaming into a dedicated InfluxDB time-series store, with AI/ML optimization feeding calibration back to operators through a React console.
Result
3x faster analytical queries and 40% less manual calibration time.
Role
Software Engineer
Andritz · May 2024 - Present
Stack
.NET
Kafka
InfluxDB
SQL Server
React
Python
Azure
What I did
Cut manual calibration time 40% with AI/ML industrial optimization systems
Led migration to InfluxDB, improving time-series query performance 3x
Built .NET and SQL Server backends with React front ends for teams in Germany and Canada, working in English daily
Standardized the team's AI-assisted pipeline: custom agents, hooks, MCP servers, automated review and Playwright E2E
Event-driven integrations with Apache Kafka; Python for automation and data scripting