What the work actually is

This is a build-and-ship backend role, not a labelling or model-critique engagement. The core deliverable is an automated pipeline that ingests BIL files originating from PeopleSoft, classifies and reconciles charges (PO matching, charge code splits), applies configurable business rules rather than hard-coded conditionals, and pushes results into the Impact billing system. Day to day you are writing Python (or Go) services deployed on Cloud Run, wiring Cloud Scheduler and Pub/Sub for ingestion, designing PostgreSQL schemas on Cloud SQL with audit logging, data versioning and batch rollback, and standing up Cloud Build test gates, Cloud Monitoring alerts and Cloud Armor controls. The AI-integration element is scoped and secondary: calling LLM services (Vertex AI SDK) from backend code where classification or extraction benefits from it, with deterministic rules and audit trails still carrying the financial correctness burden.

What the screen looks for

  • Production GCP depth, not exam recall. Expect follow-ups on Cloud Run cold starts and concurrency, Pub/Sub delivery semantics and idempotency, and Cloud SQL connection handling — named services alone won't survive a second question.
  • Financial-data seriousness. Reconciliation, accounts payable flows, and ERP integration (PeopleSoft, SAP, Oracle) come up specifically; the screener probes how you handled partial failures, restatements, and reconciliation breaks where money was at stake.
  • Rules-engine judgment. Why a configurable framework instead of code branches, how rules are versioned and tested, who owns rule changes after go-live.
  • PostgreSQL detail. Schema design, indexing choices, transaction boundaries, and how you implement rollback of an entire billing batch without losing the audit record.

Logistics

Full-time engagement, located at the Hyderabad office — this is not listed as remote or async, and candidates outside commuting distance should treat location as a real gate. Experience band is 5–8 years. Reporting line is the Solution Architect / Tech Lead, and the role extends through UAT into go-live support, so expect release-window pressure and production on-call reality near cutover. Pay is undisclosed on the platform listing; compensation is set in the client conversation after screening, so ask early rather than inferring a band.