Artificial Intelligence (AI) Software Engineer

General Dynamics Mission Systems, Inc•,

1d•$142,696 - $158,303•Remote

About The Position

General Dynamics Mission Systems (GDMS) engineers a diverse portfolio of high technology solutions, products and services that enable customers to successfully execute missions across all domains of operation. With a global team of 12,000+ top professionals, we partner with the best in industry to expand the bounds of innovation in the defense and scientific arenas. Given the nature of our work and who we are, we value trust, honesty, alignment and transparency. We offer highly competitive benefits and pride ourselves in being a great place to work with a shared sense of purpose. You will also enjoy a flexible work environment where contributions are recognized and rewarded. If who we are and what we do resonates with you, we invite you to join our high-performance team! This role involves building production AI systems for a 10,000-person enterprise, not prototypes, demos, or proofs of concept. You will work in a multi-model environment (Claude, Codex, Gemini, open source) on real enterprise problems such as legacy modernization, ERP replacement, and manufacturing intelligence. AI-assisted development is the default workflow, meaning you will use AI to build AI. The team is small, the problems are hard, and your code ships to production, directly changing how a major defense enterprise operates.

Requirements

Bachelor's degree in Software Engineering, or related Science, Technology, Engineering or Mathematics field, plus a minimum of 8 years of relevant experience; or Master's degree, plus 6 years relevant experience.
Ability to obtain a Department of Defense Secret security clearance is required at time of hire.
Applicants selected will be subject to a U.S. Government security investigation and must meet eligibility requirements for access to classified information.
Due to the nature of work performed within our facilities, U.S. citizenship is required.
Production experience building applications with LLM APIs — you have deployed generative AI services that real users relied on, not just experimented with in notebooks
Strong Python development skills — you write clean, testable, production-grade code, not scripts
Experience with RAG pipelines, vector databases, and document ingestion workflows in production environments
Experience building and consuming REST APIs — you have integrated AI services with enterprise systems and data platforms
Containerized deployment experience — Docker, Kubernetes, CI/CD pipelines. You have shipped code through automated pipelines, not manual deployments.

Nice To Haves

Experience with agent frameworks — LangChain, LangGraph, or similar tools for building multi-step, tool-using AI workflows
Experience with multiple cloud platforms (AWS, Azure, GCP) including cloud-native AI services
Hands-on use of AI-assisted development tools (Claude Code, GitHub Copilot, Cursor) as part of your daily workflow
Experience with streaming data pipelines (Kafka, Airflow) and production data infrastructure
Model monitoring and evaluation — you have built systems to track AI service reliability, not just accuracy metrics in a notebook
Commercial technology background — SaaS, healthcare, fintech, or platform engineering. Defense experience is not required.

Responsibilities

Production AI services. Build and deploy agentic workflows, RAG pipelines, and LLM-integrated applications using Python, LangChain/LangGraph, and commercial foundation models (Claude, Codex, Gemini, open source).
Data-to-insight pipelines. Implement document ingestion workflows that transform unstructured enterprise data into structured models for AI reasoning — embeddings, vectorization, knowledge graphs.
API integration. Design and build secure API interfaces that connect AI services to internal tools, enterprise platforms (Oracle, IFS, Snowflake, PLM, MES, CRM), and data sources.
Deployment and reliability. Containerize and deploy AI services using Docker and Kubernetes. Build monitoring and evaluation pipelines to track model reliability, latency, and operational performance.
Prompt engineering at scale. Design, test, and optimize prompts and agent configurations for production use — not demos.
Build things that work. Your default response to a problem is code, not a document.
Ship AI systems that real users depended on in production.
Take a problem statement and figure out the right approach, comfortable working without detailed specs.
Care about reliability as much as capability — monitor what you deploy.
Move fast without being reckless. Know when to iterate and when to get it right the first time.