Fariborz Bagherzadeh

• Deployed a Python model to Azure App Service, enabling real-time inference via a public endpoint. • Built a streaming ETL pipeline (2 sources → InfluxDB) for reliable, analytics-ready time-series storage. • Unified 20+ datasets into a reusable pipeline, speeding up experiments and transfer-learning workflows. • Reduced MAE by 50% by building an adaptive LSTM with drift detection for non-stationary data. • Improved forecasting NRMSE by 4.9% using occupancy-aware LSTM/Seq2Seq features on energy data.

• Delivered dashboard-driven web apps with clear KPI definitions and maintainable implementations. • Built backend services and API integrations to improve the consistency of event/data capture. • Designed MongoDB schemas to support fast queries, aggregations, and operational reporting. • Used Git branching and reviews to keep changes traceable and reduce regressions. • Automated build/test/deploy steps with CI/CD-style workflows to shorten release cycles.