AI Full Stack Developer

Position Summary

Precise Software Solutions Inc. is seeking an AI Full Stack Developer to join a forward-deployed, hands-on delivery team supporting the government’s Enterprise AI Solution Delivery and Support Services program. You will design, build, test, and operate secure, Government-owned AI-enabled applications, including agentic AI components built within the government’s AI agent security framework, and play a lead role in testing, evaluation, and AI-specific security testing to ensure every release is accurate, safe, and production-ready.

Location: Washington, DC metro area (hybrid) 3 days a week on site in DC

Citizenship: U.S. citizen or green card holder/EAD

Background Investigation: Eligible for public trust clearance- Government background investigation required

Key Responsibilities

• Design, develop, and enhance AI-enabled full-stack applications as part of a small, forward-deployed team, from accessible front-end interfaces (Section 508 / WCAG 2.1 A and AA) to back-end APIs, services, and workflow automation.
• Build Large Language Model (LLM) integrations, Retrieval-Augmented Generation (RAG) pipelines, vector search, and agentic AI workflows using Model Context Protocol (MCP)-based tools, engineered to the government’s AI agent security framework (OWASP Top 10 for LLM Applications, NIST AI RMF) with prompt engineering, retrieval quality controls, and human-in-the-loop oversight.
• Plan and execute functional, performance, and AI-specific adversarial testing (prompt injection and jailbreaks), evaluating outputs for accuracy, hallucination, bias, and safety; build and maintain CI/CD and MLOps pipelines with security scans and Test, Evaluation, Verification and Validation (TEVV) gates, and prepare release artifacts.
• Monitor production AI systems, perform root-cause analysis, support incident response and scheduled helpdesk support, and contribute to Responsible AI documentation (model cards, impact assessments); participate in two-week Agile sprints using Government-furnished AI engineering tools (e.g., GitHub Copilot, with model inference via Azure AI Foundry, Google Vertex AI, or AWS Bedrock), taking part in code and prompt reviews under the same review and security gates as hand-written code, and keeping code, prompts, and documentation in Government-approved repositories with full traceability.