Hamzy Chohan

Owns the architecture of a solar telemetry platform covering 148 sites and roughly 8M meter and inverter readings a day, setting layer boundaries across Delta tables on ADLS Gen2, Polars transforms into a Fabric Lakehouse, and dbt models in a Fabric Warehouse. Designed the per-site execution model using serverless Azure Container Apps jobs launched from Dagster, and led the June 2026 portfolio-wid

Recommended Azure Container Apps over AKS and Microsoft Fabric over Databricks, then built REST API ingestion from AlsoEnergy PowerTrack into partitioned Delta tables on ADLS Gen2 for 148 sites. Reduced peak transform memory by 75% using Polars with a 7-day rolling window and partition-pruned Delta MERGE, and modeled the Fabric Warehouse in dbt with SCD Type 2 dimensions.

Owned the analytics data platform for a ride-hailing operation covering 40K+ rides a day, running containerized ETL on AWS EKS into BigQuery, MongoDB Atlas, and SQL Server/SSAS. Landed 5M+ driver GPS events a day in MySQL and moved them onward with Debezium CDC into Kafka, and built dimensional models for rides, drivers, and finance.

Ran the Apache Airflow 2.x platform on Google Kubernetes Engine, operating 90+ production DAGs across parcel, species, grid, and flood data for a 48-state SaaS product. Built nationwide parcel ingestion from Regrid for 3,100+ US counties into PostGIS, and cut Buildable Area runtime from 9 hours to 3.5 hours by moving multi-state raster work to PySpark on Databricks.

Built the product usage analytics layer in BigQuery that became the reporting source of truth for enterprise accounts, replacing a manual weekly reporting cycle. Automated source availability and freshness monitoring across 10+ external data sources and published 40+ Mapbox tilesets from PostGIS and GDAL rasters.

Bachelor's degree in Finance from Baruch College, CUNY.