AI Software Developer (Mid-level) - TS/SCI with Polygraph

Description

The company builds and integrates AI-enabled features into mission applications that run on classified networks.

You will work under the Senior AI Engineer on machine learning and large language model features, within set architecture patterns and governance rules, alongside general software engineering tasks.

The role focuses on production-quality delivery in environments where AI models must run locally with no external API access, and where security, auditability and data handling are everyday concerns.

Responsibilities

• Build and maintain software that includes AI and ML features inside air-gapped and accredited government cloud environments.
• Implement LLM features such as retrieval-augmented generation, embeddings, summarization and classification using locally hosted models and program-approved patterns.
• Build and maintain data pipelines for AI features, handling classification correctly and keeping data clean.
• Connect AI capabilities to existing program systems and workflows through REST APIs and structured data formats.
• Write clean, testable, well-documented code, take part in code reviews and follow team engineering standards.
• Add the required logging, audit and evaluation instrumentation under the program's AI rules of engagement.
• Support deployment and monitoring of AI-enabled systems, and keep improving them, including bringing model updates into the environment.
• Work with program staff and stakeholders to turn requirements into working software.

Requirements

• Active TS/SCI clearance with a polygraph, or one in process.
• Bachelor's degree or equivalent experience.
• At least 2 years of professional software development experience.
• Strong Python skills and working knowledge of at least one other language, such as JavaScript/TypeScript or Java.
• Hands-on experience with AI/ML in real projects, covering model inference, embeddings, vector search or prompt engineering.
• Experience with REST APIs, JSON, SQL and Git.
• Ability to work in restricted or disconnected development environments, or proven ability to adapt to them.
• Strong problem-solving and communication skills, and the ability to work full time on-site in a SCIF.
• Preferred: experience running or integrating locally hosted LLMs such as Llama, Mistral, Hugging Face transformers, vLLM or Ollama, rather than only commercial APIs.
• Preferred: familiarity with PyTorch or similar ML frameworks.
• Preferred: experience with Docker and CI/CD pipelines, especially in offline or artifact-mirrored setups.
• Preferred: exposure to government cloud such as Azure Government or AWS GovCloud, or to IC or DoD program environments.
• Preferred: understanding of data security, classification handling and responsible AI practices.
• Preferred: prior NRO or IC program experience.

Conditions

Full-time, 40 hours per week, fully onsite in Chantilly, Virginia.

No travel required.

US citizenship required.

Mid-level role, requiring 2+ years of related experience.

Identity verification uses biometrics and AI. You must be on camera for virtual interviews, and your biometric data is used for verification and security.

The company is an equal opportunity employer that welcomes individuals with disabilities and protected veterans.

Benefits

• Medical plan options, some with Health Savings Accounts, plus dental and vision plans.
• 401(k) with pre-tax and post-tax contributions up to IRS limits and a company match.
• Full flex work weeks where possible, and paid time off including vacation, sick and personal time, holidays, paid parental, military, bereavement and jury duty leave.
• Short and long-term disability, life, accidental death and dismemberment, personal accident, critical illness and business travel accident insurance, provided or available.
• An AI-powered career tool that suggests career steps and learning, and an internal mobility team.
• Likely salary range of $140,250 to $189,750, set by experience, location and possibly contract terms, and may fall outside this range.