Data Analyst Intern HQ - San Francisco - Data

CareerPuck Inc; San Francisco (CA); On-site; USD 25,000 - 30,000; Part time

Benefits offered by this job

Hybrid work culture

Job summary

Puck is seeking a data-focused role to own acquisition and enrichment of a licensing dataset end to end, building the foundation for our AI-driven screening platform. You’ll work across ingestion, matching, and quality control to turn disparate sources into a coherent dataset.

This role blends entrepreneurial thinking with rigorous analysis and requires comfort handling government data requests, large joins, and cost-conscious modeling.

Qualifications

• Proficiency in Python and Pandas or SQL with large data joins.
• Strong spreadsheet modeling skills with defensible unit economics.
• High tolerance for administrative grind and persistence with government agencies.
• Clear written communication for drafting records requests and findings.
• Ability to judge data quality and flag questionable fields rather than shipping them.

Responsibilities

• File public records requests: draft, follow up, track status.
• Manage payments and fulfillment for records from states.
• Maintain a source registry with provenance, refresh cadence, licenses, and data quality issues.
• Build the unit economics of the dataset and cost per enriched record by source.
• Model enrichment ROI and give spend recommendations.
• Design programmatic pages for the data set and track organic traffic.

Skills

Python

Pandas

SQL

Spreadsheet modeling

Administrative grind

Written clarity

Data quality judgment

Job description

Puck is a company that moves beyond job descriptions to highlight the teams behind open roles. Our AI-powered talent attraction and screening platform that helps companies hire better candidates, faster. By combining authentic talent branding with AI-driven candidate screening, Puck helps talent teams widen top-of-funnel and move the right people through the process, efficiently.

Puck is assembling is a clean, current, national picture of licensed clinicians — who they are, where they practice, and how to reach them.That data does not exist in one place. Someone has go get it, make it usable, and figure out what it's worth.

That's this role. You'll own the acquisition and enrichment of our licensing dataset end to end, and then help us answer the commercial questions that follow. This is a build-the-foundation job, not a maintain-the-dashboard job. It’s a mix of an entrepreneurial, analytical and technical role.

What you'll do

• File public records requests where data isn't published: draft the request, follow up, track status, escalat when needed
• Handle payment and fulfillment logistics for states that charge for records
• Maintain a source registry documenting provenance, refresh cadence, license terms, and known quality issues for every source

Enrich

• Build and run a matching waterfall, in cost and confidence order
• Handle entity resolution and dedupe across sources with different name conventions, stale addresses, and partial identifiers
• Measure enrichment quality honestly.

Model

• Build the unit economics of the dataset: cost per enriched record by source and waterfall stage, marginal cost of each incremental quality tier
• Model a giveaway strategy and stress-test it against churn and conversion assumptions
• Recommend where enrichment spend has positive ROI and where it doesn't

Distribute

• Design programmatic page related to our data set
• Do the keyword research behind them, build the page templates, and track what actually ranks
• Report on organic traffic and conversion from these pages

What we're looking for

Required

• Python and pandas, or SQL, at a level where a 200k-row join with dirty keys isn't intimidating
• Real spreadsheet modeling skill — you can build a unit-economics model with assumptions broken out and defensible
• High tolerance for administrative grind. A meaningful share of this job is polite persistence with government agencies.
• Written clarity. You'll be drafting records requests and writing up findings for a small team that will act on them.
• Judgment about data quality. We want someone who flags “this field looks filled but I don't trust it” rather than shipping it.

How we work

• No heavyweight orchestration frameworks unless there's a clear reason.
• Everything documented as you go. The source registry is a deliverable, not an afterthought.
• Hybrid work culture (1 day / week in the office).

$20 hourly