We build high-throughput crawling and structured data extraction infra designed for AI agents and LLM pipelines.
If you’ve built scrapers recently, you already know the state of the market:
- Traditional proxy providers (Bright Data, Oxylabs, Smartproxy, Proxybase) give you raw IPs or massive unparsed HTML blobs, leaving you to deal with DOM bloat, token waste, and brittle parsing pipelines.
- Modern tools like Firecrawl or Jina made clean Markdown extraction easy for hobby setups, but they often struggle when you throw heavy SPAs, high-concurrency jobs, or hardened anti-bot walls (Cloudflare Under-Attack / Turnstile, Kasada, DataDome) at them.
- Rolling your own Playwright / Puppeteer cluster works until target sites deploy new JA4/TLS fingerprints, and your data team ends up spending 30% of their sprints fixing broken crawlers instead of shipping product.
We're tackling this by combining a low-latency Rust fast-fetch engine (JA4/TLS and HTTP/2 frame emulation) with an isolated stealth browser fleet, zero-code schema validation (JSON Schema/Pydantic), and native Model Context Protocol (MCP) support.
Trade-offs / where we fit: If you just need to curl a few static Wikipedia pages, a 10-line Python script is better and cheaper. But if you need reliable, structured data out of bot-defended sites without babying headless browsers, that's what we do.
We're looking for:
1. Rust Systems Engineer: High-concurrency network I/O, Axum/Tokio, TLS/BoringSSL fingerprinting, proxy routing matrices, and low-latency data pipelines.
2. Browser Security Engineer: Deep familiarity with browser internals (Firefox/Chromium source), canvas/WebGL/audio fingerprint evasion, CDP/WebDriver stealth, and familarity with anti-bot defenses.
Our stack: Rust, Next.js 16/React 19, TypeScript, hardened Firefox/dwx_crawler, Linux/Xvfb, SQLite WAL/Postgres.

Rust Systems Engineer, Browser Security at Datamint