Verified Job
Work from home
Research Engineer - Technical Writer
Work from home
#731664
Remote / WFH
Aurora
Job Description
Research Engineer — AI Search Methodology
Remote · Full-time or contract
$60K–$110K USD + equity
The company
Content engineering platform for AI-mediated discovery across ChatGPT, Perplexity, Gemini, Claude, and AI Overviews.
Brands use it to understand where they appear, why they are cited, and which content influences AI-generated answers.
Customers include Ramp, Chime, Carta, Webflow, Klaviyo, Kayak, and Rippling. Investors include Greylock, Unusual Ventures, Wing VC, and Founder Collective. Revenue grew 5x last year.
The role
You will help build the research function for a new category: how AI search systems retrieve, rank, cite, summarize, and recommend content.
This is not copywriting or product documentation. It is an applied research role with a public writing surface.
You will design studies, build datasets, analyze messy AI-search data, validate findings, and turn them into public research that technical readers, SEOs, analysts, journalists, marketers, and customers can trust.
Best fit: research engineer, data scientist, search/relevance analyst, computational journalist, or technical SEO analyst with strong Python/SQL and clear writing.
The problem
AI search is early, noisy, and easy to misread. Most market claims rely on small samples, unclear prompts, biased datasets, cherry-picked examples, and confident conclusions from fragile evidence.
Good research means knowing what counts as retrieved vs. cited, which confounders matter, and where the evidence stops. The hard part is making a claim that survives skeptical readers.
What you’ll own
• Research design: prompt sets, query taxonomies, samples, comparison groups, and measurable outcomes.
• Dataset construction: AI responses, cited URLs, retrieved URLs, fan-out queries, SERP results, scraped pages, extracted text, schema, freshness, and domain-level features.
• Analysis: SQL, spreadsheets, and Python to inspect outliers, segment results, compare cited vs. retrieved-but-not-cited pages, and separate signal from noise.
• Methodology: filters, exclusions, normalization rules, sensitivity checks, limitations, and defensible claims.
• Validation: sample size, duplication, prompt effects, model variability, skewed domains, alternative explanations, and measurement error.
• Writing: publish clear reports where the methodology matters as much as the headline.
Who this is for
• 2+ years in data science, research engineering, applied research, search/relevance analysis, SEO analytics, data journalism, or technical content with data.
• Strong SQL and practical Python, especially pandas.
• Statistical judgment around sample size, confounders, measurement error, correlation vs. causation, and reproducibility.
• Comfort with messy web data: URLs, redirects, scraped HTML, extracted text, metadata, page types, headings, and domains.
• Interest in AI search behavior: retrieval, ranking, citations, query expansion, fan-out, snippets, embeddings, and model variation.
• Clear written English and strong taste for precision.
Nice to have: SEO, AEO/GEO, LLM visibility, scraping, browser automation, HTML parsing, data visualization, regression, classification models, or published analytical work.
Logistics
$60K–$110K USD · Equity for full-time employees · Fully remote · Full-time or contract
About Aurora
Aurora helps exceptional technical and analytical talent find the right role at ambitious startups worldwide.
Remote · Full-time or contract
$60K–$110K USD + equity
The company
Content engineering platform for AI-mediated discovery across ChatGPT, Perplexity, Gemini, Claude, and AI Overviews.
Brands use it to understand where they appear, why they are cited, and which content influences AI-generated answers.
Customers include Ramp, Chime, Carta, Webflow, Klaviyo, Kayak, and Rippling. Investors include Greylock, Unusual Ventures, Wing VC, and Founder Collective. Revenue grew 5x last year.
The role
You will help build the research function for a new category: how AI search systems retrieve, rank, cite, summarize, and recommend content.
This is not copywriting or product documentation. It is an applied research role with a public writing surface.
You will design studies, build datasets, analyze messy AI-search data, validate findings, and turn them into public research that technical readers, SEOs, analysts, journalists, marketers, and customers can trust.
Best fit: research engineer, data scientist, search/relevance analyst, computational journalist, or technical SEO analyst with strong Python/SQL and clear writing.
The problem
AI search is early, noisy, and easy to misread. Most market claims rely on small samples, unclear prompts, biased datasets, cherry-picked examples, and confident conclusions from fragile evidence.
Good research means knowing what counts as retrieved vs. cited, which confounders matter, and where the evidence stops. The hard part is making a claim that survives skeptical readers.
What you’ll own
• Research design: prompt sets, query taxonomies, samples, comparison groups, and measurable outcomes.
• Dataset construction: AI responses, cited URLs, retrieved URLs, fan-out queries, SERP results, scraped pages, extracted text, schema, freshness, and domain-level features.
• Analysis: SQL, spreadsheets, and Python to inspect outliers, segment results, compare cited vs. retrieved-but-not-cited pages, and separate signal from noise.
• Methodology: filters, exclusions, normalization rules, sensitivity checks, limitations, and defensible claims.
• Validation: sample size, duplication, prompt effects, model variability, skewed domains, alternative explanations, and measurement error.
• Writing: publish clear reports where the methodology matters as much as the headline.
Who this is for
• 2+ years in data science, research engineering, applied research, search/relevance analysis, SEO analytics, data journalism, or technical content with data.
• Strong SQL and practical Python, especially pandas.
• Statistical judgment around sample size, confounders, measurement error, correlation vs. causation, and reproducibility.
• Comfort with messy web data: URLs, redirects, scraped HTML, extracted text, metadata, page types, headings, and domains.
• Interest in AI search behavior: retrieval, ranking, citations, query expansion, fan-out, snippets, embeddings, and model variation.
• Clear written English and strong taste for precision.
Nice to have: SEO, AEO/GEO, LLM visibility, scraping, browser automation, HTML parsing, data visualization, regression, classification models, or published analytical work.
Logistics
$60K–$110K USD · Equity for full-time employees · Fully remote · Full-time or contract
About Aurora
Aurora helps exceptional technical and analytical talent find the right role at ambitious startups worldwide.
