Data Engineer, Lead Data Cleanup and Verification (Fintech / Business Lending)

About us

Fundmate is a fast-growing business funding company in Brooklyn. We have millions of business leads and years of real bank statement data. We need one sharp data person to turn that into a small, verified list of high-revenue businesses our sales team can trust.

The mission

Cut a database of millions of leads down to the best 200,000–300,000 businesses: real, operating, high-revenue, reachable, and clean. Then build a system that filters every new lead list automatically, so only the good ones get through.

What you’ll do

• Audit and dedupe millions of records across AWS (RDS/Postgres, S3) and Salesforce. Match the same business across different names, phones and emails.
• Validate every phone number: active, mobile vs. landline, carrier, disconnected. Remove bad and wrong numbers.
• Remove anything that isn’t a real business. Verify businesses using public records (Secretary of State filings, business registries).
• Pull real monthly revenue from years of bank statement PDFs, using AI document extraction (LLMs or OCR) at scale.
• Remove defaults, blacklisted merchants, low-revenue businesses and opt-outs. Flag Do Not Call numbers.
• Score and rank every business on revenue, history and reachability.
• Deliver the verified list within the first 30 days.
• Build an automated pipeline: when we upload a new vendor list, it gets cleaned, scored, and only the qualified leads go to sales.
• Keep the list fresh every month.

What you need

• Strong SQL and Python (pandas or similar), working with millions of rows
• AWS experience: RDS, S3, Lambda
• Real experience with record matching and deduplication
• Experience with phone and data validation APIs
• Experience using LLM APIs to pull data from documents (PDFs, statements)
• Bonus: Salesforce API, B2B lead data, lending or fintech
• You move fast and ship. We want results in weeks, not months.

Job details

Full-time. Open to contract-to-hire. On-site or hybrid in Brooklyn, NY 11213.

Starting at $120,000/yr.

How to apply

Send a short note about the largest messy dataset you cleaned up: how many records, what you did, and what the result was. Finalists do a paid 48-hour test on a sample of our data.

Pay: From $120,000.00 per year

Application Question(s):

• What’s the largest dataset you’ve deduplicated, and how did you match records?
• Which phone validation tools have you used?
• Have you pulled data from PDFs using AI or OCR? At what scale?
• Can you start within 2 weeks?

Work Location

Hybrid remote in Brooklyn, NY 11213