×
Register Here to Apply for Jobs or Post Jobs. X

Chief Architect

Job in New York, New York County, New York, 10261, USA
Listing for: TEKsystems c/o Allegis Group
Full Time position
Listed on 2026-07-30
Job specializations:
  • Software Development
    Software Architect
Job Description & How to Apply Below
Location: New York

Description
We are seeking an Enterprise Architect that can take our existing platform and modernize it to meet the current and future demands of the Secondary Mortgage market. This person will be hands on technical but also leading small team of developers and QA.
1. Platform Architecture Governance
Own and enforce the five-layer architectural blueprint across all systems, data, and integration layers. The canonical data model is non-negotiable infrastructure. Establish and maintain schema governance - no uncontrolled field creation, no custom code bypassing platform tools. Every field has a defined source tier, business definition, and provenance record.
  • Complete and govern the DynamoDB-to-Postgres migration, including schema design, data validation, and deprecation sequencing for legacy DynamoDB tables.
  • Own the Silver data dictionary architecture: field definitions, source lineage from stated data through manufactured data to authoritative Silver designation, tier classification (Tier 1 direct institutional pull, Tier 2 third-party, Tier 3 document submission), and target/consumer documentation.
  • Establish and chair the Architecture Review Council. All significant design decisions are reviewed, documented, and held to pattern standards before implementation.
  • Define and enforce the Anti-Rube Goldberg principle: one governed way to solve each class of problem. Reuse over novelty. No proliferation of bespoke solutions.
2. Rules Engine and BPM Deployment
Deploy and govern the componentized rules engine and BPM layer that is the operational core of the certified loan process across all three seller channels:
Flow/Delegated, Non-Delegated/Mini-Corr, and Bulk/Institutional. This layer is what makes the 600-to-6 complexity abstraction possible at scale.
  • Complete the Camunda (or equivalent) rules engine evaluation, make the selection decision, and lead deployment. The rules engine must be business-authorable, engineer-reviewed, and vendor-extractable.
  • Enforce the workflow/rules separation principle in all builds: workflow orchestrates, rules validate. Logic never embeds in workflow steps.
  • Build the componentized rules library in collaboration with the credit risk and underwriting leadership, supporting channel-specific configurations without duplicating core logic.
  • Design the rules engine to support the eight certified loan process gate validations across delegated, non-delegated, and bulk channels, with third-party data integration at each gate where applicable.
3. AI Execution Layer and Prompt Engineering Governance
Establish and govern the AI execution layer as the fifth layer of the platform architecture. AI generates ETL mappings, rules components, test cases, and workflow artifacts - but only within governed guardrails. The business owns outcomes. Engineering owns the platform. The Enterprise Architect owns the governance model that makes this safe.
  • Establish the Center of Excellence (CoE) for AI-assisted development: prompt libraries, pattern standards, testing/retraining protocols, and senior review gates for all AI-generated artifacts before production deployment.
  • Own prompt engineering governance centrally. One owner, one library, one set of standards. No proliferation of individual prompt sets outside the governed framework.
  • Lead adoption of AWS Q for Developer, Git Hub Copilot, and Anthropic Claude in engineering workflows, with measurable targets for code generation, test automation, and documentation quality.
  • Build and govern the LLM-assisted rules generation capability: ingesting underwriting guidelines (including the Fannie Mae Seller Guide and investor-specific overlays), generating componentized rules, and validating outputs against the canonical data model.
  • Design the AI co-pilot model for the non-delegated manufacturing process: reviewing loan files during origination, identifying exceptions, and preparing clean handoffs to buyers - eliminating post-closing defect discovery.
4. ETL, Integration, and Third-Party Data Architecture
Own the integration architecture that connects our canonical data model to the third-party data sources that validate and elevate stated data across the loan manufacturing…
To View & Apply for jobs on this site that accept applications from your location or country, tap the button below to make a Search.
(If this job is in fact in your jurisdiction, then you may be using a Proxy or VPN to access this site, and to progress further, you should change your connectivity to another mobile device or PC).
 
 
 
Search for further Jobs Here:
(Try combinations for better Results! Or enter less keywords for broader Results)
Location
Increase/decrease your Search Radius (miles)
0
200
Filters
Education Level
Experience Level (years)
Posted in last:
Salary