About this role.
ICON seeks an experienced AI Software Engineer to design and ship production agentic-AI capabilities for its construction technology platform and government technology programs. The role centers on LLM-powered products, multi-agent orchestration, tool use, MCP integrations, RAG pipelines, and rigorous evaluation practices. Candidates need at least six years of software engineering experience, strong Python and/or TypeScript skills, and demonstrated experience shipping LLM-based systems. This is a remote full-time role with periodic project-based travel to ICON headquarters in Austin, Texas, and the selected candidate must be able to obtain and maintain a security clearance.
Role DNA
A quick view of the complexity, pace, ownership and collaboration implied by the job description.
Job Complexity
5/5Pace & Pressure
4/5Autonomy Level
5/5Communication Load
4/5Salary analysis
Estimated compensation compared with the broader US market for similar roles.
Core skills
Skills and capabilities most closely associated with this opportunity.
Sample interview questions
I would explain the workflow's user goal, tools, orchestration logic, guardrails, and deployment architecture. I would also describe measurable evaluation criteria, fallback paths, observability, and the production outcomes such as reduced manual effort or improved task completion.
I would begin with source governance and ingestion, then use document-aware chunking, metadata enrichment, embeddings, and a vector store with hybrid retrieval where appropriate. I would add reranking, citation requirements, access controls, and offline evaluation sets to validate answer quality, grounding, and retrieval performance.
I would define task-level success criteria and test both individual agents and end-to-end workflows with representative scenarios. I would track tool-call correctness, groundedness, latency, cost, failure modes, and human-review outcomes, then use those findings to refine prompts, tools, routing, and guardrails.
I would describe creating or integrating secure tool interfaces with clear schemas, authentication, least-privilege permissions, validation, and auditability. I would emphasize making tool results predictable for models and implementing robust error handling, timeouts, and safe retries.
I would use a controlled experimentation environment with approved models, sanitized data, logging controls, and clear release gates. Before production deployment, I would perform threat modeling, validate data handling and access controls, establish monitoring and incident procedures, and involve security stakeholders early in the design process.
ICON is looking for a pragmatic builder who has made the leap from traditional software engineering into the emerging world of agentic AI development. As an AI Software Engineer at ICON, you will be responsible for designing and shipping AI-powered tools, workflows, and autonomous agents that accelerate ICON’s construction technology platform. This role reports to the Sr. Director, Defense Technology Programs and is available for remote work with periodic travel to ICON HQ in Austin, TX as projects require.
RESPONSIBILITIES
Design and build agentic AI systems that automate complex, multi-step workflows across ICON’s software platform
Develop LLM-powered features and products using state-of-the-art foundation models and APIs (e.g. Anthropic, OpenAI)
Architect and implement multi-agent pipelines, tool-use systems, and Model Context Protocol (MCP) integrations
Build robust RAG systems including document ingestion, chunking strategies, embedding pipelines, and vector retrieval
Collaborate with software and domain teams to identify high-leverage AI automation opportunities and translate them into shipped products
Own the full development lifecycle of AI features: prototyping, evaluation, deployment, and iteration
Serve as a technical resource and informal mentor on agentic AI best practices across the engineering organization
Stay at the leading edge of the agentic AI landscape and bring emerging techniques into production
MINIMUM QUALIFICATIONS
6+ years of professional software engineering experience with a strong foundation in backend or full-stack development
Demonstrated experience building and shipping production-grade products using LLMs and agentic frameworks
Proficiency in TypeScript and/or Python
Deep understanding of prompt engineering, context management, and LLM reasoning patterns
Experience with tool-use, function calling, and agent orchestration (e.g. LangChain, LlamaIndex, Claude Code, or custom implementations)
Ability to design and execute structured evals to measure and improve AI system performance
Strong communication skills and comfort working cross-functionally with both technical teams and non-technical stakeholders
- Must be able to obtain and maintain security clearance.
PREFERRED SKILLS AND EXPERIENCE
Experience with MCP (Model Context Protocol) server development and integration
Familiarity with vector databases (e.g. pgvector, Pinecone, Weaviate)
Experience with fine-tuning, RLHF, or model evaluation pipelines
Background in ML fundamentals: embeddings, transformers, attention mechanisms
Exposure to structured output generation and LLM-based data extraction
Experience working with code generation agents (e.g. Claude Code, Cursor, Devin-style systems)
Familiarity with AWS and serverless infrastructure for AI workloads
Experience working in or alongside government, defense, or regulated environments
Interest or background in the AEC (architecture, engineering, construction) industry
Rate Range: $126,400 – $180,000
ICON is an equal opportunity employer committed to fostering an innovative, inclusive, diverse and discrimination-free work environment. Employment with ICON is based on merit, competence, and qualifications. It is our policy to administer all personnel actions, including recruiting, hiring, training, and promoting employees, without regard to race, color, religion, gender, sexual orientation, gender identity, national origin or ancestry, age, disability, marital status, veteran status, or any other legally protected classification in accordance with applicable federal and state laws. Consistent with the obligations of these laws, ICON will make reasonable accommodations for qualified individuals with disabilities.
Furthermore, as a federal government contractor, the Company maintains an affirmative action program which furthers its commitment and complies with recordkeeping and reporting requirements under certain federal civil rights laws and regulations, including Executive Order 11246, Section 503 of the Rehabilitation Act of 1973 (as amended) and the Vietnam Era Veterans’ Readjustment Assistance Act of 1974 (as amended).
Headhunters and recruitment agencies may not submit candidates through this application. ICON does not accept unsolicited headhunter and agency submissions for candidates and will not pay fees to any third-party agency without a prior agreement with ICON.
As part of our compliance with these obligations, the Company invites you to voluntarily self-identify as set forth below. Provision of such information is entirely voluntary and a decision to provide or not provide such information will not have any effect on your employment or subject you to any adverse treatment. Any and all information provided will be considered confidential, will be kept separate from your application and/or personnel file, and will only be used in accordance with applicable laws, orders and regulations, including those that require the information to be summarized and reported to the federal government for civil rights enforcement purposes.
Upload your resume now
- To unlock remote work opportunities and be discovered by global employers.
This job listing has been manually reviewed by the Jobicy Trust & Safety Team for compliance with our posting guidelines, including verification of the company's legitimacy, accuracy of job details, clarity of remote work policy, and absence of misleading or fraudulent content.
