About the Opportunity

Our client is an early-stage technology company building an AI-powered financial intelligence platform for startup founders. The platform combines financial data, analytics, and AI to help founders understand their business and make informed decisions.

We are looking for an AI Engineer to own the intelligence layer of the product, including the AI analyst, background agents, and evaluation systems that ensure AI outputs are accurate, reliable, and traceable.

This is an opportunity to join a small, founder-led engineering team as an early AI hire. You'll work directly with the CTO and founding team, take ownership of key AI initiatives, and make independent technical decisions from day one.

You don't need to be a finance expert, but you should care deeply about accuracy, reliability, security, and responsible AI, particularly when working with financial information.

What You'll Do

LLM & AI Application Engineering

• Build and own the AI analyst experience, including agentic chat loops, tool selection, grounding, citations, refusal behavior, and streaming.
• Develop production LLM applications that interact with users across product and connected platforms.
• Design and extend typed, tenant-scoped tools that provide reliable and traceable data to AI systems.
• Build structured AI workflows with appropriate guardrails and deterministic data sources.

AI Agents & Automation

• Design and build bounded AI agents that can plan, gather information, process data, and provide outputs for user approval.
• Develop and improve agent orchestration frameworks and workflows.
• Determine when to use AI models versus deterministic functions to achieve reliable outcomes.

LLM Evaluation & Quality

• Build and maintain AI evaluation programs, including golden datasets, chat evaluations, regression testing, and drift monitoring.
• Measure model and prompt performance before and after changes.
• Establish quality standards for AI-generated responses and ensure features are validated before production release.

Model & Prompt Operations

• Manage model and prompt configurations, registries, and versioning.
• Optimize prompt caching, token usage, model selection, and AI costs.
• Establish per-tenant limits and monitor cost per conversation.

AI Observability

• Instrument AI interactions and track prompts, tools, tokens, cache performance, latency, and cost.
• Use observability data to identify reliability, performance, and cost improvements.

AI Innovation

• Explore and implement practical AI capabilities such as:
  • Document and spreadsheet intelligence
  • Meeting intelligence
  • Retrieval over connected data
  • AI-powered data ingestion
  • Emerging LLM capabilities

AI Security & Data Protection

• Implement PII protection and redaction before information reaches AI models.
• Build safeguards against prompt injection and jailbreak attempts.
• Maintain strong tenant isolation and secure access controls.
• Design AI features with security and data privacy as core requirements.

Product & Customer Collaboration

• Participate in customer conversations to understand how users interact with financial data and AI.
• Translate customer feedback into improvements that can be rapidly tested and shipped.
• Work closely with the founding team to identify and prioritize high-impact AI opportunities.

Who We're Looking For

Required Skills & Experience

• Production experience building LLM/Generative AI applications used by real users.
• Hands-on experience with AI agents, agent loops, and tool/function calling.
• Experience with structured outputs, streaming, grounding, and/or RAG.
• Strong TypeScript development experience.
• Experience with Next.js and Postgres.
• Strong SQL and data-handling skills.
• Experience building or maintaining LLM evaluation and testing frameworks.
• Understanding of AI reliability, hallucination mitigation, and model failure modes.
• Strong security awareness when working with sensitive or customer data.
• Experience using AI-assisted development tools such as Claude Code, Cursor, GitHub Copilot, or similar.
• Ability to work independently, make technical decisions, and operate effectively in an early-stage environment.
• Strong product mindset and a bias toward shipping high-quality solutions.

Nice to Have

• Experience with Anthropic/Claude or OpenAI APIs
• Experience with Vercel AI SDK
• Experience with MCP
• Experience with AI observability and tracing
• Experience with LLM evaluation platforms or automated AI testing
• Experience with financial data, fintech, accounting, or SaaS products
• Python experience for AI evaluation, experimentation, or analysis

Technology Stack

Languages: TypeScript, Python
Frontend: React, Next.js
AI/LLM: Anthropic API, Claude, OpenAI, Vercel AI SDK
AI Systems: Agents, Tool Calling, Structured Outputs, RAG, MCP
Database: PostgreSQL, SQL
AI Development: Claude Code, Cursor, GitHub Copilot or similar
Focus Areas: LLM Evaluation, AI Observability, Prompt Engineering, AI Security, Agentic Workflows

Working Time Zone: Pacific Time