Automation / Data Engineer (Part-Time) — Build a Lead Enrichment Engine

We run a B2B lead generation pipeline built on public business filing data. Records come in on an automated pull, get scored and tiered, and route to our sales team through CRM.

The problem: contact enrichment is still manual. Two people research each record by hand to find phone numbers and email s, and they can't keep pace with volume. Our highest-value records go stale before anyone touches them.

Your job is to make that step disappear.

This is a build project first, with ongoing part-time work maintaining and extending the system afterward if it goes well.

What you'll build

Automated enrichment. Integrate third-party data APIs to resolve contact information without human involvement. Includes deduplication and caching so we never pay twice for the same lookup, phone line type validation, and fallback logic when a lookup fails.

Automated lead scoring. Our raw data contains signals we aren't currently computing. You'll derive them at ingest and build them into a scoring model that runs on every record automatically. I need to be able to adjust scoring weights myself without asking you for a code change.

Prioritized review queue. Records that fail automated lookup go to a human queue, ordered by record value and freshness rather than first in first out.

Monitoring and alerting. This matters as much as the build. The system runs unattended against paid APIs and feeds a queue our sales team depends on. I need failure alerts, spend anomaly alerts, rate limit handling with backoff, and retry logic that doesn't duplicate paid lookups. A silent failure costs real money.

Documentation. A deliverable, not an afterthought. System documentation, a runbook for common failures, and a walkthrough where I can demonstrate I understand how to operate it. I built the current pipeline myself and I'm not willing to end up unable to troubleshoot my own system.

Required
Strong Python, or n8n plus Python
Real experience integrating third-party APIs, including handling rate limits, pagination, retries, and auth
Experience working with large volumes of messy record data: deduplication, normalization, fuzzy matching on names and addresses
A database background (PostgreSQL or similar)
You've built something that ran unattended in production and you know what it takes to keep it running
Clear English

Nice to have
Google Apps Script and Google Sheets API
Close CRM or comparable CRM APIs
Experience with skip trace, identity, or B2B contact enrichment vendors
Dashboarding (Looker Studio, Metabase, or similar)
Any background in lending, MCA, fintech.