Enterprise Data Architect

Programmers.io; New York (NY); Hybrid; USD 120,000 - 180,000; Full time

Job summary

Programmers.io is seeking a seasoned data architecture professional for a contract/full-time role in New York City, offering hybrid or remote work. You will design and implement the fraud data platform and the end-to-end data mesh, focusing on real-time decisioning, lakehouse modeling, and data virtualization to empower investigators with contextual insights.

You will participate in design reviews, write specs, and set partitioning, latency budgets, and state-management strategies.

Qualifications

• Experience implementing data mesh architecture with Palantir Foundry in a fraud/financial services context.
• Strong understanding of real-time data pipelines and architecture patterns.
• Ability to define data products, governance, and federation in a distributed environment.

Responsibilities

• Fraud data architecture: own end-to-end design across latency tiers: decision plane, investigation plane, and model/analytics plane.
• Working knowledge of Palantir Data Foundry and Data Mesh implementation; lakehouse design and build.
• Data mesh implementation: define topology, contracts, quality metrics, and federated governance.
• Data virtualization: design federated access layer for investigation and analytics.
• Real-time data path: design streaming ingestion, ISO 20022 event model, and low-latency feature state store.
• Non-functional design: latency budgets, throughput modelling, resilience, data residency, retention, lineage, and encryption.

Skills

Palantir Foundry

Data Mesh architecture

Real-time data path

Latency budgets

Job description

Location: New York City, NY (Hybrid or Remote)

Hire Type: Contract / Full Time

Primary Skills: Data Mesh Architecture implementation using Palantir Foundry

Secondary Skills: Financial Services, Fraud Management platform build experience is a great addition

Design and build the data architecture that fraud detection runs on: the real-time decisioning path, the lakehouse that models are trained from, the data products the estate is organized around, and the virtualization layer that gives investigators context without proliferating copies.

This is a builder's role. The successful candidate will be in design reviews, writing and reviewing technical specifications, and making calls on partitioning, latency budgets and state management. Candidates whose recent experience is exclusively advisory will not be a fit.

Key responsibilities

Fraud data architecture- Own the end-to-end design across three latency tiers: the decision plane (sub-150ms, materialised state, no federated access), the investigation plane (seconds, virtualized), and the model and analytics plane (lakehouse)

Working knowledge of Palantir Data Foundry and Palantir's Data Mesh implementation experience Lakehouse design and build.

Data mesh implementation - Define the fraud data product topology, the contract standard between producing domains and the fraud platform, the quality metrics attached to each product, and the federated governance model that keeps ownership distributed while evidence stays central

Data virtualization - Design the federated access layer for investigation and analytics

Real-time data path - Design the streaming ingestion, the ISO 20022 canonical event model, the bounded-window enrichment join between payment and channel telemetry, and the low-latency feature state store

Non-functional design - Latency budgets, throughput and capacity modelling, resilience and failure posture, data residency, retention, lineage and encryption