Sreekanth Potlabathini·Jun 13, 2025Clustering Multidimensional Time-Series Sensor Data — A Friendly Walk-Through of KL, JS, EMD, KS…1. Why compare distributions at all?
Sreekanth Potlabathini·Jun 10, 2025Predicting Battery Life — A Practical Guide with Production-Ready ML model“A dead sensor means at least 24 hours of blind security — unacceptable when your cabin hides a ton of gold.”
Sreekanth Potlabathini·Jun 10, 2025Listening to Machines: A Practical Guide to Sound‑Based Anomaly Detection with LSTMTL;DR We added a low‑cost microphone into a sensor attached to to a conveyor, stream 16 kHz audio to Python, convert each 3-seconds slice…
Sreekanth Potlabathini·Jun 10, 2025How a Straight-Forward Rule-Based Model Beat a “Real” Machine-Learning Pipeline in Production.“Start with the simplest thing that could possibly work.” — Kent Beck
Sreekanth Potlabathini·Jun 10, 2025Built an Anomaly Detection Model Without Any Ground Truth — Just Feedback“We didn’t have any labels. But we still built a model that learned to spot problems — just by listening, watching, and adapting.”