Web Crawling - Research Engineer

Thinking Machines Lab•San Francisco, CA

23d•$350,000 - $475,000•Onsite

About The Position

We're hiring a Software Engineer to build and own our web-crawling systems, from distributed collection at internet scale through filtering, deduplication, and deciding what data we keep. The ideal candidate has built and scaled a web crawler or large-scale data-acquisition systems. In this role, you'll write and own production systems: the crawler itself, the infrastructure that runs it at scale, and the pipelines that turn raw crawls into usable pretraining data. You'll work closely with our pretraining and data teams to understand what's actually moving model quality, but this is fundamentally an engineering role, not a research one.

Requirements

8+ years designing, building, and scaling web crawlers, scrapers, or large-scale distributed data-acquisition systems
A track record of owning crawler or data-acquisition infrastructure at internet scale
Strong software engineering skills in a language such as Python, Go, or Rust, with real experience in distributed systems
Working knowledge of the practical and legal considerations of large-scale web data collection (robots.txt, rate limiting, licensing)

Nice To Haves

Experience applying machine learning to crawl selection, extraction, or data quality classification at internet scale
Experience setting technical direction for a crawling, data acquisition, or search infrastructure team, whether or not that was your formal title
Experience designing systems for petabyte-scale storage and processing
Track record of open-source contributions to crawling, scraping, or data infrastructure tools
Background at a search engine (crawling, indexing) or a frontier AI lab's data acquisition team

Responsibilities

Design and scale the web crawler and ingestion infrastructure that sources Inkling's pretraining data
Build pipelines for large-scale extraction, deduplication, and data quality filtering
Build specialized crawlers for high-value or hard-to-reach data sources
Work with the pretraining team to understand how changes in crawled data affect model performance
Improve the reliability and efficiency of crawling and ingestion infrastructure at petabyte scale
Help set technical direction for this area as it grows, and bring other engineers up to speed on what you've learned