# Senior Integration Engineer at Spherecast

- Company: Spherecast
- What the company does: We support multi-channel e-com brands with demand planning, replenishment and inventory optimization to avoid stockouts & overstock and improve inventory visibility. Backed by Y Combinator.
- Company website: https://www.spherecast.ai/
- Type: Startups (AI role)
- Level: Senior
- Location: New York City
- Work setup: On-site
- Posted: 2026-09-23
- Apply by: 2026-11-07
- Apply: https://jobs.ashbyhq.com/spherecast/b2a93cab-1c4f-499e-91a6-0488acfebe75
- Page: https://www.1752.vc/careers/jobs/spherecast-senior-integration-engineer/

## About the role

Data is the lifeblood of operating an excellent supply chain. Without clean data, everything crashes: forecasts drift, orders land wrong, and trust evaporates. The Spherecast Platform is the real-time operating system for global consumer-goods supply chains, and Agnes , our AI Supply Chain Manager, runs on top of it. The platform and Agnes are only as good as the data flowing in and the actions flowing back out. That's where you come in.

## What they're looking for

- Hands-on experience with our backend stack: Python, FastAPI, SQLAlchemy, and Pydantic, with a track record of shipping production data pipelines and integrations
- A strong instinct for pipeline design: fetch, validate, normalize, upsert, built idempotent and re-runnable, with clean schema boundaries between the remote system's world and ours
- Proficiency in SQL, writing high-performance queries and bulk upserts despite billions of data points for even a single customer
- Respect for the write path: pushing AI-approved actions into a customer's ERP demands safety rails, dry runs, failure capture, and auditability
- A habit of building for operability: structured logging, sync-run tracking, and alerting (we use PagerDuty) so failures are caught, triaged, and attributed, including telling an external stakeholder when the error is on their side
- Claude Code, Codex, and OpenCode are your best friends. You know how to parallelize instances to make sure we're shipping 24/7, even while you sleep, and you know the ins and outs of each of them at any time

Tags: Engineering