A well-funded, early-stage AI company building an intelligence layer for institutional finance is hiring an Applied AI Engineer to own the full data substrate behind its research agent, already in daily use by dozens of financial institutions. This is a hands-on, full-stack role covering everything from document ingestion through to the search design the agent relies on to decide what it's even allowed to know. Based on-site in New York, this role is for someone who wants to go deep across the entire pipeline rather than own a single layer of it.
Company Culture
- On-site in New York, working closely with the founding team
- Founded by a small team with strong finance and technical research backgrounds
- Product used in live, high-stakes work by both sell-side and buy-side firms
- Backed by a well-known syndicate of financial and venture investors
- Small, senior engineering team where this hire sets the definition of done for the core data layer
What's on Offer
- Salary: up to $220k, dependent on experience
- Direct ownership of a foundational piece of the product, not a single feature
- Seat close to the founders on a system already in production use at scale
What You'll Be Doing
- Owning the connector layer ingesting from common enterprise sources (cloud storage, email, partner data feeds), with event-based and scheduled sync
- Running a parsing pipeline across external vendors with fallback-chain orchestration, preserving block-level source provenance for citation
- Building an LLM metadata-extraction stage with multi-model fallback chains, feeding deterministic entity resolution
- Indexing into a modern search/data store under a strict, filterable metadata contract, and shaping the identification layer around how the agent plans multi-step work
- Running containerised workloads on cloud infrastructure with GitOps-based deployment, and defining infrastructure as code
- Hardening the internal skill/tool system so external systems can consume it as APIs, and building customer-driven R&D on the core data layer
- Making improvements measurable through extraction-quality evals and search precision/recall against labeled query sets
What We're Looking For
Core Technical Requirements
- Strong Python experience across modern async web frameworks and ORMs
- Experience with at least one search or data store technology (e.g. OpenSearch, Elasticsearch, PostgreSQL)
- Comfort with asynchronous Python
Infrastructure & Stack
- A genuine willingness to go down the stack, from application code into infrastructure
- Familiarity with Kubernetes and infrastructure-as-code practices is a strong plus
AI-Native Practice
- Exposure to multi-model LLM extraction chains and structured outputs
- Some backend TypeScript experience is a plus
- Active use of coding agents (e.g. Claude Code, Cursor, Codex) as part of daily practice
If this sounds like your kind of role, feel free to reach out for a confidential chat.