Artificial Intelligence Engineer

Role Type: Full-time | Location: Onsite | Stack: Python

The Company We are building a vertically integrated, AI-native operating platform for institutional real estate investment. The platform automates and connects every stage of the investment lifecycle — from acquisition through renovation, leasing, and refinance — replacing manual processes with code, AI, and SaaS integrations. Remote-first, no on-ground employees.

How We Build Software We follow an automation-first hierarchy: code first, then specialized AI, then LLMs, then offshore humans, then US contractors. Humans are the last resort, not the first option. All feature work follows a spec-driven development approach. You write specs, orchestrate Claude Code agents, review output, and iterate. You are an engineer who directs AI, not a typist who writes every line by hand. If you think what took an hour still takes an hour or AI coding tools are a gimmick, this is not the role.

• Build and deploy AI agents and automation workflows for real estate operations (underwriting, document processing, lead qualification, contractor coordination, etc.)
• Design multi-step agent architectures with tool-calling, structured outputs, retries, fallbacks, and human-in-the-loop escalation
• Build Python microservices that expose AI capabilities via REST APIs to the main platform
• Implement cost management and token optimization across all LLM usage
• Convert manual/human tasks into AI-automated workflows based on detailed process specs
• Work with the product team to identify which of our 178 mapped tasks can be shifted from Human to AI
• Monitor, evaluate, and improve agent performance with metrics and logging

• 5+ years professional Python backend engineering (not scripting or notebooks)
• Production experience with LLM APIs (OpenAI, Anthropic, etc.) — not demos or tutorials
• Experience with agent frameworks (LangChain, CrewAI, AutoGen, or custom)
• Strong understanding of structured outputs, function/tool calling, and prompt engineering for deterministic behavior
• Experience designing and deploying microservices with REST APIs
• Ability to build robust error handling, retry logic, and fallback patterns for non-deterministic AI systems
• Cost-aware: experience managing token budgets and optimizing LLM spend
• Comfortable with remote async work across time zones
• Spec-driven mindset: can take a detailed process spec and translate it into working automation

• Experience with real estate, proptech, or financial services automation
• Familiarity with RAG (Retrieval Augmented Generation) pipelines
• Experience with document parsing/extraction (OCR, PDF processing)
• Knowledge of vector databases (Pinecone, Weaviate, pgvector)
• Experience with Claude Code or similar AI-assisted development workflows

• A fullstack/frontend role — you will not build UI or work on the Next.js platform
• A data science or ML training role — we use pre-trained LLMs, not custom models
• A no-code automation role — this is real engineering, not Zapier workflows
• A research role — you ship production systems, not papers