What our engineers are learning as they ship — data quality, anomaly detection, AI reliability, cloud, and security.
NeuAlto co-founder Hemanth K Rajasekhar on two decades at Wipro, why DeltaMax and OptiMax exist, what he looks for when hiring, and what AI actually changes about engineering work.
No posts in this topic yet — check back soon.
Shorter notes, published there first.
Building DeltaMax taught us that catching every anomaly doesn't solve the problem — it creates alert fatigue. The breakthrough was context: a Trust Score and an Agent Summary that explain how severe a change actually is, rather than just flagging that something moved. It's also why we shifted from rule-based checks to learned patterns; you can't write a rule for a failure you haven't imagined yet.
AI is only ever as reliable as the data it learns from. Good anomaly detection does more than spot outliers — it recognizes patterns, surfaces meaningful change, and separates genuine risk from opportunity. We look at where IQR, KNN, Isolation Forest, and ARIMA each earn their place in a production stack.
