At INTENT I worked on the backend platform powering an agricultural analytics SaaS product serving 63 enterprise clients across 1.7 million acres.
Some of the work I'm most proud of:
Geospatial data pipelines — Designed and operated large-scale raster processing pipelines handling several TB of elevation data across the continental US using GDAL, Cloud-Optimized GeoTIFFs, and VRTs on AWS and GCS. This work fed terrain feature computation that supported ML-based crop prediction models.
Job orchestration system — We were initially using Celery for compute-heavy jobs but hit memory exhaustion issues specific to how GDAL parses GeoTIFFs under load. I diagnosed the problem and built a lightweight DAG-based job orchestrator in Python and PostgreSQL that gave us better control over job lifecycle and resource behavior. It was later picked up by teammates for unrelated workloads across the platform.
AI pipeline — Built a production system that batched prompt execution across 100+ agronomic trials against the Gemini API, storing structured model outputs for downstream analytics and application use.
I also owned backend architecture across Python services, PostgreSQL/PostGIS data models, and AWS/Kubernetes infrastructure, and contributed across the full stack including React-based frontend mapping applications.
