Three engineers from Detroit. Zero outside funding.
We met through the SMRP Detroit chapter. Two of us came off plant floors. One of us built ML systems for industrial applications. We all believed the same thing: plant maintenance teams shouldn't need a data scientist to use data that's already flowing through their historians.
The people building Gearcadence
A small team with deep manufacturing operations experience, building precisely the tool we wish had existed during our plant floor years.
Lukas Reinhardt
CEO & Co-Founder
Eight years as a reliability engineer at a Detroit-area Tier 1 automotive supplier. Lukas built the Python anomaly-detection prototype that became Gearcadence after watching the same $180,000 gearbox fail catastrophically three times in four years — each failure preceded by a vibration signature visible in the SCADA historian that nobody had the tools to catch in time. His focus is product strategy, early customer deployments, and making sure every feature survives contact with an actual plant floor.
Tomasz Kovalski
CTO & Co-Founder
Former reliability engineer turned systems architect, with six years in condition monitoring and IIoT infrastructure at a Midwest industrial equipment OEM. Tomasz designed the Gearcadence edge gateway architecture — including the real-time anomaly detection pipeline that runs at 10kHz without cloud dependency. He leads all hardware and software engineering, and owns the OPC-UA and CMMS integration layer. His rule: if it doesn't work when the plant Wi-Fi drops, it doesn't ship.
Adaeze Nwosu
Head of Machine Learning
Applied ML engineer with a background in time-series anomaly detection and industrial sensor data — previously building predictive models for energy grid equipment at a utility-scale infrastructure company. Adaeze owns the degradation modeling framework that powers Gearcadence's time-to-failure forecasting. She built the calibration workflow that reduces the baseline training window to two weeks per asset, and designed the confidence interval methodology used for TTF output. Her standard: every model result must be explainable to a maintenance supervisor without a statistics background.
We met at an SMRP Detroit chapter meeting in 2022. Within an hour, all three of us had described the same problem from different vantage points.
Lukas had the plant floor frustration. Tomasz had the hardware architecture instinct. Adaeze had the ML framework. Gearcadence was incorporated in 2023, self-funded through consulting contracts while we built the product we always needed.
Talk to an engineer, not a sales rep
Every demo is run by a member of the core team. We'll spend 45 minutes on your equipment, your historian, and your actual maintenance challenges — not a feature tour.
Request a Demo