Research Engineer (Focused on Search/IR)

Research Engineer (Focused on Search/IR)

You'll own and advance the search and information retrieval systems at the core of Firecrawl - the infrastructure that determines how we find, rank, index, and serve web content at scale. This is a hands-on, full-stack search role where you'll build and operate everything from ingestion pipelines to serving layers. If you've built search indexes at massive scale and care deeply about ranking quality, freshness, and retrieval speed, this is the role.

Salary Range: $180,000-$270,000/year (Range shown is for U.S.-based employees. Compensation outside the U.S. is adjusted fairly based on your country's cost of living. You can explore how we calculate this here: https://www.firecrawl.dev/careers/compensation.)

Equity Range: Up to 0.15%

Location: San Francisco, CA or Remote (Americas, UTC-3 to UTC-10)

Job Type: Full-Time

Experience: 3+ years building search/IR systems at scale

Visa: US Citizenship/Visa required for SF; N/A for Remote

About Firecrawl

Firecrawl is the easiest way to extract data from the web. Developers use us to reliably convert URLs into LLM-ready markdown or structured data with a single API call. In just a year, we've hit 8 figures in ARR and 90k+ GitHub stars by building the fastest way for developers to get LLM-ready data.

We're a small, fast-moving, technical team building essential infrastructure super-intelligence will use to gather data on the web. We ship fast and deep.

What You'll Do

• Build and operate search indexes at massive scale: Design, build, and maintain the indexing infrastructure that powers Firecrawl's core product. You'll handle billions of documents and care about every millisecond of latency and every byte of storage.
• Own the full stack from ingestion to serving: You don't just build one piece - you own the entire pipeline. Ingestion, processing, indexing, ranking, query understanding, and serving. When something breaks at 3am, you know where to look because you built it.
• Solve ranking, relevance, and query understanding: Make sure the right content surfaces for the right queries. You'll build and iterate on ranking models, relevance scoring, and query parsing systems that directly impact product quality.
• Tackle freshness, dedup, and incremental indexing: The web changes constantly. You'll build systems that keep our index fresh without re-crawling everything, deduplicate content intelligently, and handle incremental updates at scale without rebuilding from scratch.
• Run experiments and ship results to production: You design experiments, measure results rigorously, and ship winners to production fast. You don't need someone to tell you what to try next - you have a backlog of ideas and the judgment to prioritize them.
• Collaborate with the research team: Work closely with the Head of Research and the RL-focused Research Engineer to connect search/IR improvements with model training and broader product strategy.

What We're Looking For

Someone who has built search indexes at massive scale. Not a tutorial project - real indexes serving real traffic with real latency requirements. You've dealt with the hard problems: sharding strategies, index compaction, schema evolution, and the operational complexity of keeping billions of documents queryable and fast.

Hands-on with ranking, relevance, and query understanding. You've built or meaningfully improved ranking systems. You understand BM25, learned ranking, embedding-based retrieval, and when to use which. You can reason about relevance tradeoffs and you've shipped ranking changes that moved metrics in production.

Owns the full stack: ingestion → index → serving. You're not a specialist who only touches one layer. You've built and operated the entire search pipeline - from how documents enter the system to how results get served. You understand the dependencies between layers and you make good architectural decisions because you see the whole picture.

Has solved freshness, dedup, and incremental indexing problems. You know that building the initial index is the easy part. Keeping it accurate, fresh, and deduplicated at scale is where the real engineering lives. You've built systems that handle continuous updates without full rebuilds and you've debugged the subtle correctness issues that come with incremental processing.

Self-directed experimenter who ships without handholding. You generate your own hypotheses, design your own experiments, and ship your own code. You don't wait for a roadmap or a sprint planning meeting. You see what needs to improve, you try something, you measure it, and you ship it if it works.

Backgrounds that tend to do well: Search engineers at companies with large-scale indexes (web search, e-commerce search, document search). IR researchers who've shipped their work to production. Infrastructure engineers who've built and operated real-time indexing pipelines. Engineers from Elasticsearch, Algolia, Vespa, or similar search infrastructure