SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Baselayer is rebuilding the identity and verification layer for financial institutions across the US. The company has built the most complete business graph in America, fusing public records, IRS data, sanctions lists, web signals, and fraud telemetry from 2,200+ financial institutions to resolve any business and its associated humans in milliseconds. With 98% match rates achieved in under two years (vs. legacy bureaus at 60% over 50 years), Baselayer is trusted by over 20% of US financial institutions including FIS, Rho, and Socure.
You'll join a small, high-ownership team solving real-time entity resolution at unprecedented scale. The role focuses on building LLM-driven agents that answer questions loan applications don't ask: what a business actually does, where it lives on the web, whether named individuals match public records, and whether open-web signals contradict the application narrative. This is production data infrastructure, not research.
Key responsibilities include: owning industry/category classification from heterogeneous signals; building discovery and verification systems for real web presence (filtering aggregators, parked domains, impersonators); linking individuals to businesses via public evidence; developing risk/legitimacy scoring from web signals; and building shared agent infrastructure including provider-agnostic base agents, tool registries, eval harnesses, and instrumentation for token/tool tracing.
You'll own model selection, agent design, prompt and tool engineering, eval methodology, and cost control across your enrichment surface. The technical depth is real: graph embeddings, fraud propagation models, sub-100ms latency traversal, and expanding identity verification beyond finance into gig platforms, marketplaces, and commerce infrastructure.
Minimum requirements: shipped LLM agents to production (not demos); strong async Python with structured-data libraries and relational databases; experience across multiple frontier LLM providers and at least one agent framework; built or maintained eval methodology with golden datasets and scoring functions; browser automation experience including headless browsers and anti-bot evasion; informed opinions on structured-output reliability (JSON-schema vs. function calling vs. extractors).
What sets you apart: web scraping at scale with anti-bot evasion and residential proxies; eval-framework experience (LangSmith, Braintrust, Evals); entity resolution/record linkage/fuzzy matching at scale; devtools-protocol browser automation; tool registry abstraction over multiple LLM providers; cost/latency optimization expertise.
