🔧 AI Platform Architect – Green‑Field Real‑Estate ML Spin‑Off (NYC‑Hybrid) 🔧
Imagine you had Zillow‑scale data, carte blanche to build an AI platform from scratch, and a war chest bankrolled by a profitable parent company. Welcome to our stealth spin‑off: a five‑person “research lab” that’s about to weaponize ML for every homeowner, investor, and contractor in America.
We’re hunting for a hands‑on Platform Architect who can design, build, and scale the entire ML stack—data ingestion to real‑time inference—while shaping the engineering culture from day one.
What You’ll Own
•Platform blueprint: choose the cloud primitives (we lean AWS + Bedrock), craft the data lakehouse, stand up feature stores, and script CI/CD for models and micro‑services.
•End‑to‑end pipelines: automate ingestion of terabytes of listing, loan, and renovation data; orchestrate training and continuous evaluation; push lightning‑fast, explainable inference APIs.
•Infra‑as‑code + observability: Terraform, CDK, Grafana/Prometheus—whatever keeps shipping reliable and rollback painless.
•Security & scale: multitenant data‑governance, SOC‑friendly controls, and an architecture that survives “hacker‑news‑front‑page” traffic spikes.
•Tech leadership: mentor a lean squad of ML engineers, collaborate with our Product Lead on roadmap trade‑offs, and report directly to a PhD‑level Head of Data.
The Stack (starting point)
AWS (EKS, Bedrock, SageMaker, S3, DynamoDB) • Python/TypeScript • Airflow or Prefect • Terraform/CDK • Vector DBs (PGVector, Pinecone, or similar) • Docker • GitHub Actions • Whatever best‑in‑class tools you champion next
Must‑Haves
•5+ yrs architecting production ML/AI platforms (not PoCs) at meaningful scale.
•3+ end‑to‑end pipelines shipped—raw data → trained model → live traffic.
•Comfort with LLM/RAG patterns, embeddings, and high‑volume inference.
•Infra‑as‑code disciple; you can spin an environment from zero without clicking consoles.
•Communication chops to explain latency graphs to execs and kernel panics to juniors—often in the same hour.
Nice‑to‑Haves
•Prop‑tech, fintech, or mapping/geo experience.
•Buy vs. build scars—you know when to grab a managed service and when to roll your own.
•Startup stint or founder DNA; you thrive on white‑boards, post‑its, and shipping before lunch.
Why You’ll Love It
•Hybrid sweet‑spot: small Chelsea/West Midtown loft, 2-3 days in office for the high‑bandwidth magic
•Stable upside: parent company profitability + meaningful equity in the spin‑off.
•Comp Package: $160 K – $250 K base (DOE) + equity + flexible cash bonus. Visa sponsorship for exceptional talent.
•Day‑one impact: first architect on the ground—you set the patterns everyone else will follow.