Solutions Architect Principal

General Dynamics IT; United States; On-site; USD 150,000 - 210,000; Full time

Job summary

General Dynamics IT in Albany, NY, is seeking a Solutions Architect Principal to design and build agentic AI systems for the NY MMIS program, with strict response-time, availability, and HIPAA/HITECH compliance.

You will lead architectural direction for AI governance, tool interfaces, and reference implementations, mentor engineers, and ensure security, cost, and reliability of AI workloads within a regulated environment.

Qualifications

• Education: Bachelor's in Computer Science, Software Engineering, Mathematics, or a related technical field; an advanced degree in Computer Science, Machine Learning, or Artificial Intelligence preferred
• Experience: 8+ years of professional software engineering or systems architecture experience, including 1+ year building and operating production large language model systems, with at least one agentic system taken to production and supported afterward

Responsibilities

• Architect, build, and operate multi-agent systems - planning, tool use, memory, and reflection loops - that run as governed, auditable, and reproducible steps inside the delivery pipeline
• Design and implement the tool and context layer that agents depend on: Model Context Protocol servers, tool and output schemas, retrieval pipelines, and context engineering standards
• Build the evaluation harness before the feature - eval sets, scoring rubrics, and regression suites that prove a non-deterministic system improved rather than assert it
• Own performance, cost, and reliability for AI workloads: tracing, token and spend budgets, model selection and routing, fallback behavior, and caching
• Establish the guardrail and safety architecture - PHI stage gates, prompt injection and data exfiltration defenses, least-privilege tool scoping, and human-in-the-loop approval checkpoints
• Evaluate models, frameworks, and vendors on evidence: run bake-offs, publish results honestly, name lock-in and failure modes, and recommend against adoption when warranted
• Produce the architecture others build from - reference architectures, interface requirement documents, design patterns, and decision records - and defend it in architecture review boards
• Raise the engineering organization's ceiling through pairing, design review, and the standards governing how the program applies AI to protected health information

Skills

Python

TypeScript

Java

Go

Rust

System architecture

Education

Bachelor's in CS/Software Eng/Math

Advanced degree in CS/ML/AI preferred

Tools

Amazon Bedrock

Anthropic Claude

Claude Code

pgvector

OpenSearch

Terraform

Terragrunt

GitLab CI/CD

AWS

Job description

Type of Requisition: Regular

Clearance Level Must Currently Possess: None

Clearance Level Must Be Able to Obtain: None

Public Trust/Other Required: None

Job Family: Technical Solutions

Job Qualifications: Skills: AI Agent Security, Solution Architecture, Systems Engineering Certifications: None Experience: 8 + years of related experience US Citizenship Required: No

SOLUTIONS ARCHITECT PRINCIPAL

Support mission-critical initiatives and enable the growth of our business as a Solutions Architect Principal at GDIT. Here, you'll become an integral part of how GDIT is able to ensure the safety and security of our nation.

MEANINGFUL WORK AND PERSONAL IMPACT

As a Solutions Architect Senior Advisor, the work you'll do at GDIT will be impactful to the mission of the New York State Department of Health Medicaid Management Information System (MMIS) program. You will play a crucial role in designing and building the agentic AI systems that support the eMedNY enterprise at Corporate Woods in Albany, NY.

As the individual contributor for artificial intelligence and agentic engineering, you will be trusted to set the technical direction for how AI is designed, built, and governed across one of the nation's largest Medicaid claims processing environments, under extremely stringent response-time and continuous-availability service level agreements and HIPAA/HITECH and FedRAMP compliance obligations. Your influence will come from the architectures you design, the reference implementations you ship, and the engineers you mentor. In this role, your responsibilities will be:

• Architect, build, and operate multi-agent systems - planning, tool use, memory, and reflection loops - that run as governed, auditable, and reproducible steps inside the delivery pipeline
• Design and implement the tool and context layer that agents depend on: Model Context Protocol servers, tool and output schemas, retrieval pipelines, and context engineering standards
• Build the evaluation harness before the feature - eval sets, scoring rubrics, and regression suites that prove a non-deterministic system improved rather than assert it
• Own performance, cost, and reliability for AI workloads: tracing, token and spend budgets, model selection and routing, fallback behavior, and caching
• Establish the guardrail and safety architecture - PHI stage gates, prompt injection and data exfiltration defenses, least-privilege tool scoping, and human-in-the-loop approval checkpoints
• Evaluate models, frameworks, and vendors on evidence: run bake-offs, publish results honestly, name lock-in and failure modes, and recommend against adoption when warranted
• Produce the architecture others build from - reference architectures, interface requirement documents, design patterns, and decision records - and defend it in architecture review boards
• Raise the engineering organization's ceiling through pairing, design review, and the standards governing how the program applies AI to protected health information

WHAT YOU'LL NEED TO SUCCEED

Bring your initiative and drive for innovation to GDIT. The Solutions Architect Principal must have:

Education: Bachelor's in Computer Science, Software Engineering, Mathematics, or a related technical field; an advanced degree in Computer Science, Machine Learning, or Artificial Intelligence preferred

Experience: 8+ years of professional software engineering or systems architecture experience, including 1+ year building and operating production large language model systems, with at least one agentic system taken to production and supported afterward

Technical Skills

Python at expert level, plus working fluency in a second language such as TypeScript, Java, Go, or Rust

Agentic patterns and runtimes: tool and function calling, Model Context Protocol, structured output, multi-agent orchestration, durable state, and human-in-the-loop control flow

Amazon Bedrock, Anthropic Claude, and agentic coding harnesses such as Claude Code

Retrieval-augmented generation: embeddings, hybrid retrieval, reranking, and vector stores such as pgvector or Amazon OpenSearch

Evaluation and observability for non-deterministic systems; AWS cloud-native services, containers, Terraform / Terragrunt, and GitLab CI/CD

AI security and compliance: prompt injection, insecure output handling, PHI leakage prevention, HIPAA/HITECH, NIST AI RMF, and the OWASP Top 10 for LLM Applications

Role Requirements

• A computer science foundation strong enough to transfer - proven ability to pick up a framework that did not exist six months ago, read its source, judge production-readiness, and put it to work
• A demonstrated record of rapid self-directed learning; be prepared to name a technology you took from unfamiliar to production within a year and explain how
• Engineering judgment and intellectual honesty about AI - able to tell a demo from a system, quantify where a model is unreliable, and tell leadership when the answer is "not yet"
• Comfort working without an establ