Data Engineer, Ingestion Platform

Alt is unlocking the value of alternative assets, starting with the $5 B trading-card market. We let collectors buy, sell, vault, and finance their cards in one place and we are backed by leaders at Stripe, Coinbase, Seven Seven Six, and pro athletes like Tom Brady and Giannis Antetokounmpo. Our next frontier is real-time pricing at scale—the Alt Value that powers every trade, loan, and product on the platform.

The Role

Alt Value is our moat. It is only as good as the data underneath it — and that data lives on dozens of external marketplaces and auction houses that have no interest in making it easy for us to get.

That's this job. You own the pipelines that scrape, normalize, and land every transaction and listing that feeds our pricing model. When a comparable sale closes on eBay or at a major auction house, how fast it shows up in Alt Value, and whether it shows up correctly, is your call and your consequence. Pricing decisions, cash advance terms, market analytics, and every dashboard in the company sit downstream of you.

The metric you own: Data Coverage. Supporting KPIs are pipeline latency, uptime, and source coverage.

What You'll Own

The ingestion layer. Design, optimize, and own the pipelines that scrape, process, and ingest transaction and listing data from major auction houses and marketplaces. Scraper to orchestrator to warehouse, end to end.

Build for reuse first. Most of the ingestion stack — auth handling, retries, session and proxy management, monitoring — should generalize across sources. Go deep and custom only where a specific source's structure actually requires it.

Knowing before we do. Build the monitoring and alerting that tracks latency, uptime, and coverage across every source. A source going quiet should page you, not surprise a pricer three days later.

Making it cheaper and steadier. Modernize storage and processing, cut manual intervention out of the loop, and optimize for cost, performance, and reliability.

The consumers. Partner with pricing, ML, product, and analytics to deliver data clean and standardized enough that nobody downstream is writing defensive code around your output.

How You'll Use AI Here

We don't want to hear that you use Claude every day. Everyone does. We want to know what you've built with it. Concretely, in this role:

Point an agent at a site that just changed its markup and have it propose the parser fix, run it against a fixture set, and open the PR

Build a triage loop over the orchestrator — cross-reference a failed DAG against logs and warehouse state, diagnose the cause, and hand you a ranked list instead of a wall of red

Use LLM extraction for the long tail of low-volume sources where a bespoke scraper will never pay for itself

Put an MCP server over the warehouse so a coverage question is a sentence, not another one-off query

Use Claude Code or Cursor as the default way you work through a migration or a backfill, not as autocomplete

If you've built agents, written skills or MCP servers, wired up connectors, or shipped internal tooling your team actually depends on — lead with that.

What You Bring

3-4 years in data engineering or a closely adjacent seat.

Modern web scraping at scale — Playwright or comparable, plus the ability to impersonate a real browser at the network layer (TLS/request fingerprinting) when a plain HTTP client gets flagged. Selenium alone doesn't cut it anymore.

Strong Python, 3+ years hands-on. Solid SQL for analysis and transformation.

A pipeline you owned end to end in the past two years. Not a pipeline you contributed to. One that was yours when it broke.

Judgment on when to generalize and when to go bespoke — you don't rebuild the same infrastructure per source, and you don't force a one-size-fits-all approach onto a source that needs custom handling.

Orchestration in your hands, not your resume. Airflow, Dagster, or a comparable DAG system — you've built on it, been paged for it, and cleaned up after it.

Cloud development experience (AWS, GCP).

LLM-based extraction and processing at scale (e.g. LLM-as-a-Judge, Agent loops), not just UI-based Q&A and simple Skill building.