About the role

We are building an operational decision product for a multi-site industrial network — recycling yards, reconditioning shops, inventory, freight, and scrap pricing.

The people who run that network today value inbound material, choose where to cut it, move recovered components to shops, and decide whether to sell or hold scrap. Those calls are made in spreadsheets, on different clocks, with incomplete numbers. We are replacing the spreadsheet with ranked options in dollars, and a transparent reason for each ranking.

Source data is already in Snowflake. This role owns the gold layer

everything else depends on: valuation, forecasts, and a finance-facing measure of whether the recommendations paid off. If the grain is wrong, the recommendation is wrong.

What you’ll do

- Design and ship the Snowflake gold layer for the decisions above: on-hand inventory by grade and site, receipts and shipments, realised vs index prices, facility master, freight by lane, shop cost, demand and monthly commitments, and a log of recommendations accepted or overridden.
- Turn raw ERP and shop-floor extracts into tables with an explicit grain (lot, load, site × grade × month — not “whatever the source table was”).
- Build production incremental models in Snowflake (dbt or equivalent) across dev, test, and prod, with tests on uniqueness, grain, and freshness.
- Reconcile data that does not share a clean key across systems — inventory on one side, shop outcomes on the other — and document what can and cannot be joined.
- Surface missing or weak inputs (freight, capacity, operating cost) as labelled gaps. Do not drop a dimension to make a pipeline look complete.
- Work with yard ops,



shop ops, and finance so a number in the product can be traced back to a source row and survive a challenge.
- Publish data contracts for downstream forecasting and for an auditable value ledger (recommendation vs what would have happened anyway).

What we’re looking for

- 5+ years building production analytical tables used by operators or finance, not only BI dashboards.
- Expert SQL: window functions, snapshot vs transaction grain, slowly changing inventory, incremental models.
- Snowflake in production — schemas, warehouses, roles, tasks, cost-aware design.
- Dimensional or medallion modeling you have shipped for inventory, cost, or multi-site operations.
- Python for profiling, tests, and supporting SQL models.
- Evidence you have been wrong about a column, found it in real data, and corrected the model.
- Comfort in the room with operators and finance, not only with other engineers.

Nice to have

- dbt on Snowflake, with tests in the merge.
- Manufacturing, remanufacturing, recycling, or metals.
- ERP / MES data (Infor LN or similar) already landed in a warehouse.
- Commodity price or freight/lane-rate data.
- Feature tables or point-in-time snapshots for machine learning.
- Numbers that had to pass a finance or audit review.
- Gold inventory and movement spine: grade × site × time, tested.
- Incremental models from at least two source databases into gold, with freshness alerting.
- Source-to-gold map for the tables the valuation calc needs — real, assumed, and still missing.
- A written position on the hardest cross-system join: what links, at what grain, and what does not.

Skills:- Snowflake, SQL, PowerBI and Python

📌 Senior Data Engineer- Snowflake (Hyderabad)
🏢 Cutshort Lightning
📍 Hyderabad