I build intelligence systems for industrial operations.

My primary work right now is manufacturing intelligence for EV battery production. Plants have event logs, cycle times, yield data, defect records — enormous volumes of production data. What they lack is the analysis layer that connects those records to operational questions: which station is constraining the line right now, why one shift consistently underperforms another, whether a defect found at end-of-line can be traced back to where it originated. I build that layer — from raw events in MES systems through KPI computation, OEE decomposition, bottleneck analysis, and quality traceability, out to shift-level dashboards and production floor digital twin displays.

On the connected vehicle side, I lead product for a telematics and battery intelligence platform serving EV OEMs across two-wheelers, three-wheelers, four-wheelers, and commercial fleets. The hardware is automotive-grade: 4G LTE, multi-constellation GNSS, high-speed CAN for BMS and motor systems, integrated IMU, OTA firmware, IP67 housing. The software layer runs cell-level anomaly detection, battery health scoring, thermal analysis, ML-based range estimation, driver behaviour scoring, and predictive maintenance for fleet operators. The problem it solves is straightforward: OEMs and fleet operators are largely blind to what is happening inside the vehicles and batteries they have already shipped.

The industrial IoT work covers equipment condition monitoring for rotary machinery — motors, gearboxes, pumps, fans, compressors, conveyors, cranes. Edge sensors capture continuous vibration, current, and temperature signals. A cloud analytics layer runs physics-informed models on those signals to generate health scores, remaining useful life estimates, and specific maintenance recommendations — not alerts, recommendations. The shift from scheduled to condition-based maintenance is the core value: machines degrade by load and operating condition, not by the date on a service schedule.

The most technically specific product in our portfolio is a condition monitoring platform built specifically for cranes — engineered around the failure physics of heavy lifting equipment: hoist mechanics, duty cycle classification, overload events, and structural fatigue. The edge hardware runs FFT and RMS computation in real time with sub-millisecond latency. The analytics layer correlates load against mechanical stress, generates component-level health scores with confidence intervals, and produces maintenance work orders that integrate directly with CMMS platforms.