Experience | Dhia

Built an adaptive alerting system on top of the core monitoring stack (Prometheus, Thanos, Grafana) that learns each customer's normal metric behavior, detects genuine anomalies, and continuously adjusts alert limits through an automated feedback loop, replacing static, manually-set rules that led to false positives and missed incidents
Created a configuration tool for profiling metrics and calibrating these limits, which auto-generates deployment artifacts; shipped the complete solution to production Kubernetes clusters using a GitOps workflow

Developed an ETA prediction model to estimate driver travel times between two locations, replacing reliance on Google Maps
Designed and automated a full data pipeline to continuously retrain and update the model with new incoming trip data
Built an interactive dashboard to monitor real-time driver and company performance

Built and trained a machine learning model capable of detecting prohibited and high-risk items in real time
Collected, cleaned, and augmented image datasets from multiple online sources and APIs
Integrated the model into a surveillance robot