PJ Pretorius

Engineering AI product features end-to-end across FastAPI/Python services and React front-ends for enterprise clients across financial services, retail and logistics. Delivering ML-driven workflows on Microsoft Azure with Snowflake as the analytical backbone, from data modelling through to production deployment. Designing PostgreSQL schemas and REST APIs for high-volume operational data and downstream BI. Collaborating with data and platform teams to ship production AI features behind clean, user-facing UX.

Co-founded a retail-tech company building a multi-tenant operations platform for South African grocery and franchise stores; own architecture, delivery, deployment and customer support. Designed and built a FastAPI + PostgreSQL + React/Vite monorepo that runs four production apps: InvoiceFlow, MerchCommand, ShelfSense, and SupervisorHub. Implemented strict multi-tenancy (store_id isolation across every model and query), JWT/role-based authentication, audit logging and an in-app activity log.

Architected and maintained petabyte-scale data ingestion infrastructure for the MeerKAT Radio Telescope, ensuring high-throughput, fault-tolerant transfers across federated systems in South Africa's scientific ecosystem. Designed and deployed internal APIs using FastAPI to automate data flows from source to storage, enabling real-time monitoring and operational scalability. Integrated and managed distributed storage backends using iRODS, Rucio, and Ceph-backed S3, supporting mission-critical datasets for astronomical research. Wrote and maintained robust backend Python services and shell scripts performing scheduled ingest, replication, verification, and error handling.