Applied AI Engineer

Cryptio•,

9d•Hybrid

About The Position

About Cryptio Cryptio is the leading enterprise-grade financial data transformation and digital asset ERP platform for accounting, financial reporting, tokenization compliance, and loan management solutions. It is a complete back and middle-office platform that delivers financial integrity and accountability for businesses in crypto. Its platform allows users to transform on-chain and off-chain crypto activity into auditable records for accounting, treasury, tax filings, license applications, loans activity, and regulatory reporting. Having recently announced our $45m series B, now couldn't be a better time to join! What this role actually is It's a two-person engineering function: you, one other engineer, and the VP of Advisory Intelligence. You'll own complete vertical slices of the toolkit from PostgreSQL schema through API layer through React UI, shipping into production systems that clients use in live audit engagements. The role is full stack, but 70-80% of your time will be spent building out backend infrastructure - so you should be strong in this area. The catch (worth being honest about) is that low structure comes with the territory. No large team to bounce ideas off, no process that catches mistakes before they reach clients. What you get instead is real product influence, genuine autonomy, and work that's harder and more interesting than most AI engineering roles. What you'd be building Multi-step AI pipelines: prompt chaining, structured output parsing, confidence scoring, failure recovery, and audit-trail logging Full-stack feature delivery from schema to UI. You own the whole thing Financial document ingestion pipelines: parse spreadsheets and ERP exports into typed data models for downstream AI processing Background job infrastructure for long-running AI workflows An AI output evaluation framework: structured test inputs, expected outputs, scoring rubrics Stack: TypeScript, Node.js, Next.js App Router, React, PostgreSQL, AWS Bedrock / OpenAI / Anthropic SDK, Zustand, Zod.

Requirements

Shipped production full-stack TypeScript systems (not side projects).
Integrated LLMs in production.
Understand the difference between a prompt issue and an infrastructure issue because you've debugged both.
Understand PostgreSQL at a level where data modelling and indexing strategy feel natural.
Know why append-only matters in a compliance context.
Care about correctness; outputs go into audit engagements for large financial institutions.
Push back on specs when the approach is wrong.
Comfortable working async without needing a lot of external structure.

Nice To Haves

Background in finance or accounting tooling.

Responsibilities

Own complete vertical slices of the toolkit from PostgreSQL schema through API layer through React UI, shipping into production systems that clients use in live audit engagements.
Build out backend infrastructure (70-80% of time).
Build multi-step AI pipelines: prompt chaining, structured output parsing, confidence scoring, failure recovery, and audit-trail logging.
Deliver full-stack features from schema to UI.
Build financial document ingestion pipelines: parse spreadsheets and ERP exports into typed data models for downstream AI processing.
Build background job infrastructure for long-running AI workflows.
Build an AI output evaluation framework: structured test inputs, expected outputs, scoring rubrics.