AI/ML Engineer

Role Overview

Build production AI/ML for Globixs’ Real Estate vertical. You will own the Automated Valuation Model (AVM) and extend a multi-agent property intelligence pipeline, shipping features end-to-end to brokerages.

Responsibilities

• Own and improve the AVM pipeline (currently XGBoost; open to LightGBM/CatBoost or neural methods)
• Extend a multi-agent system (e.g., PropertySearchAgent, MarketAnalysisAgent, PropertyValuationAgent) using Claude and OpenAI APIs
• Integrate third-party data sources into clean, versioned feature pipelines (e.g., RapidAPI Zillow, MLS feeds, county assessor data, OpenCage, US Census)
• Design prompts, tool definitions, and agent orchestration logic running on FastAPI + Supabase + Railway
• Ship end-to-end: data ingestion → model → API → realtor-facing dashboard
• Monitor model drift, evaluate agent outputs, and build evaluation harnesses for LLM-driven components
• Collaborate with frontend engineers on Next.js integrations and work with the founder on customer-driven priorities

Requirements

• 2–5 years building ML systems in production (not just notebooks)
• Strong Python; comfortable with FastAPI, Pandas, scikit-learn, and gradient boosting (XGBoost/LightGBM)
• Hands-on experience with LLM APIs (Anthropic Claude and/or OpenAI): prompt design, tool/function calling, structured outputs
• Experience with multi-agent frameworks or orchestration patterns (e.g., LangGraph, CrewAI, OpenAI Agents SDK, or custom)
• Solid SQL and Postgres; familiarity with Supabase
• Git-based workflows, CI/CD (GitHub Actions), and cloud deploys (e.g., Railway, GCP, AWS, Vercel)
• Can read product specs, ask the right questions, and ship

Nice to Have

• Real estate/proptech domain experience (MLS, AVM, RETS, comp analysis)
• Geospatial work (e.g., PostGIS, H3, shapely, OSM)
• Demo experience with Streamlit or React
• LLM evals/observability (e.g., Langfuse, Braintrust)

Preferred Qualifications

• Client-facing delivery and cross-functional collaboration
• Strong communication and stakeholder management
• Experience owning end-to-end delivery outcomes