Spatial Front, Inc. (SFI) is seeking an AI Developer to support our growing modernization team. SFI was recently awarded the 2025 USA Today National Top Places to Work award and the 2025 Washington Post Top Workplaces. The ideal candidate will perform hands-on Python development of artificial intelligence agents and generative AI applications supporting PeopleSoft HCM and related enterprise systems.
This role is focused on building the intelligence and application behavior of production AI solutions, including agent design, prompt engineering, orchestration, retrieval-augmented generation (RAG), tool use, evaluation, and supporting Python services. The candidate will work closely with AI integration engineers, platform engineers, data engineers, functional teams, and cybersecurity personnel to develop reliable, secure, and maintainable AI capabilities for Federal Government customers.
Location
Crystal City, VA - On-Site/Hybrid
Responsibilities
- Design, develop, test, and maintain AI agents and generative AI applications using Python.
- Develop agent workflows that use large language models to interpret user requests, reason across multiple steps, select appropriate tools, retrieve information, and generate grounded responses.
- Design, test, and iteratively refine system prompts, task instructions, tool-use instructions, few-shot examples, response formats, and other prompt-engineering components.
- Apply prompt-engineering techniques to improve response accuracy, consistency, grounding, tool selection, adherence to business rules, and overall user experience.
- Develop agent orchestration logic including state management, context management, tool selection, multi-step execution, retries, exception handling, and recovery.
- Develop Python application components using boto3, botocore, and other appropriate SDKs and libraries to interact with AI, data, storage, security, and supporting cloud services.
- Integrate agents with approved tools, APIs, Model Context Protocol (MCP) services, databases, knowledge sources, and enterprise applications.
- Work with AI Integration Engineers to define tool requirements, expected inputs and outputs, validation rules, and integration behaviors needed by AI agents.
- Build and optimize retrieval-augmented generation (RAG) capabilities including query formulation, retrieval logic, context assembly, grounding, semantic search, and use of retrieved information within agent workflows.
- Develop prompt and context strategies for working with structured and unstructured enterprise information while minimizing irrelevant or unsupported model responses.
- Implement structured outputs, schema validation, response validation, guardrails, error handling, and other controls required for reliable enterprise AI behavior.
- Develop reusable Python libraries, utilities, agent components, prompts, and development patterns that can be used across multiple AI use cases.
- Create automated unit, integration, regression, and AI evaluation tests covering agent workflows, prompts, tool selection, model responses, retrieval quality, and application behavior.
- Develop and maintain evaluation methods for response quality, factual grounding, hallucination, tool-use accuracy, retrieval effectiveness, latency, consistency, and regression.
- Analyze model and agent behavior using logs, prompts, responses, tool calls, retrieved context, and downstream results to identify and correct performance or reliability issues.
- Compare and evaluate models, prompting approaches, retrieval strategies, and agent designs based on accuracy, reliability, performance, maintainability, security, and suitability for the use case.
- Collaborate with functional experts and Product Owners to translate business problems into clearly defined AI use cases, expected behaviors, acceptance criteria, and measurable outcomes.
- Support demonstrations, user testing, defect resolution, production validation, monitoring, and continuous improvement of deployed AI capabilities.
- Maintain source code, prompts, technical designs, configuration, evaluation criteria, support documentation, and other AI development artifacts.
- Participate in code reviews, architecture discussions, backlog refinement, demonstrations, testing, release readiness, and other Agile/SAFe delivery activities.
- Other duties as assigned.