Description
Summary Data Integration Engineer — Daily Unified Reporting Layer *(HubSpot, Klaviyo, SMS platform, Microsoft Dynamics 365 Business Central)* We run a B2B wholesale operation and a DTC e-commerce brand. Our customer and revenue data lives in four systems that don't talk to each other. We need one engineer to build a **read-only, once-daily ETL layer** that pulls all four into a single, deduplicated table we can report and query against. This is a data-integration / ETL project — not web dev, and not "build me an AI that runs my business." **Sources:** - **Business Central** — orders/revenue (OData/REST). The heaviest piece and our source of revenue truth. - **HubSpot** — CRM contacts/deals *plus marketing analytics* (sends, opens, clicks). Confirm you can read marketing analytics via the API before bidding. - **Klaviyo** — campaign and flow performance. - **SMS platform** — inbound/outbound message logs via API. **Scope:** - Nightly batch job pulling fresh data from all four - Lands in a single warehouse/store you recommend and justify - Identity-resolution step that collapses one customer across all four into a single record (see below) - Clean unified table + a simple way for a non-engineer to read it - Runs unattended on a schedule, with failure alerting - Handoff docs for our internal technical staff **Not** real-time, **not** two-way sync, **no** sends or writes to any source system. Data flows one direction, out. **The hard part — address this in your proposal:** the sources use different keys — Business Central on company/email, HubSpot on email, Klaviyo on email, SMS on phone number. The real work is matching a phone-keyed SMS record to an email-keyed Business Central record so the same customer isn't counted four times. In two or three sentences, tell us how you'd approach it. **Proposals that skip this will be declined.** **Access (non-negotiable):** read-only scoped credentials issued and revocable by us; no production write access; no customer data leaves infrastructure we control; signed NDA before any credential is shared. **Structure:** milestone-based, fixed scope per milestone, starting with a small paid discovery milestone. **Your proposal must include:** (1) two prior ETL/warehousing builds — not generic web/app dev; (2) your 2–3 sentence dedup answer; (3) whether you've pulled *marketing* analytics from HubSpot's API, and which endpoints/scopes; (4) the warehouse you'd recommend and one sentence why; (5) rough hour estimate per milestone. *Reference the deduplication question in your first line so we know you read this. AI-pasted boilerplate will be ignored.*