ML Software Engineer Jobs in Maryland

Leading provider of Information Operations (IO) and Irregular Warfare (IW) capabilities to the U.S. government. Integrates Information Related Capabilities (IRC) across multiple domains, environments, and geographic boundaries, supporting both domestic and overseas operation via Information Operations planners, intelligence analysts, cultural advisors, and linguists

Peraton is hiring: Generative AI Specialist / AI & ML Developer.

Location: On-site, Annapolis Junction / Jessup, MD

Required: U.S. citizenship and the ability to obtain and maintain a Top Secret/SCI with Poly security clearance.

Additional Requirement: Current, valid U.S. passport (potential OCONUS travel)

About the Role:

Peraton is seeking an experienced Generative AI Specialist to support high-impact Internal Research and Development (IRAD) initiatives focused on AI/ML-driven platforms for Operations in the Information Environment (OIE).

• This role is central to delivering advanced generative AI capabilities supporting COCOM information operations across CENTCOM, NORTHCOM, INDOPACOM, AFRICOM, and EUCOM.
• You will drive innovation in LLMs, agentic AI workflows, and generative AI application development—directly influencing Peraton’s competitive position in the defense AI landscape.

In this role, you will:

• Design and optimize generative AI solutions using GPT 4/5, Claude, Gemini, and Azure OpenAI, including prompt engineering, fine-tuning, and retrieval augmented generation (RAG).
• Implement agentic LLM architectures with structured outputs (JSON/GeoJSON), chain-of-thought/chain-of-debate methods, and multi-agent orchestration.
• Build and maintain prompt libraries, evaluation frameworks, and QA pipelines for defense-grade AI content.
• Integrate generative AI into existing platforms via documented APIs, ensuring interoperability with DoD systems (Maven, C2IE, IRIS).
• Research emerging generative AI techniques (multimodal models, synthetic data generation, AI-assisted analysis for information operations).
• Translate COCOM operational requirements into deployable AI solutions; support TTX events and platform demonstrations.
• Implement Responsible AI practices including bias detection, validation, hallucination mitigation, and HITL workflows.
• Support MVP-aligned milestone delivery, contributing to TRL progression and ROI measurement.
• Document novel prompting techniques, model configurations, and workflows that may constitute IP or trade secrets.

Minimum Qualifications:

• Bachelor’s degree in a related technical field and minimum of 5 years relevant experience. An additional 4 years of experience may substitute for the bachelor's degree.
• Minimum of 3 years in software development or data science, including 2+ years focused on generative AI, LLMs, or NLP.
• Hands-on experience with foundation models (GPT 4, Claude, Gemini, Llama, Mistral) including prompt engineering, few-shot learning, fine-tuning, and API integration.
• Strong Python skills; experience with LangChain, LlamaIndex, Hugging Face, OpenAI/Anthropic APIs.
• Experience building RAG systems using vector databases (Pinecone, Weaviate, ChromaDB, pgvector).
• Cloud experience (AWS, Azure) and deploying AI/ML models in production.
• Understanding of AI safety, hallucination mitigation, bias reduction, and prompt injection protection.
• Experience with Git, CI/CD, and DevSecOps practices.
• U.S. citizenship required.
• Ability to obtain a final TS/SCI with Poly security clearance.
• Current, valid U.S. passport for potential OCONUS travel.

Preferred Qualifications:

• Master’s degree.
• Experience with agentic AI architecture, multi-agent workflows, and LLM tool-use.
• Hands-on experience with fine-tuning, RLHF, DPO, LoRA/QLoRA.
• Familiarity with multimodal models and defense/intelligence applications.
• Experience with IRIS, OMEGA, or similar operational platforms.
• Background supporting IO, PSYOP, influence analysis, or COCOM operations.
• Knowledge of MLOps, A/B testing, and LLM evaluation frameworks (RAGAS, DeepEval).
• Experience with synthetic data generation, simulation, or Monte Carlo methods.
• Understanding of geospatial data formats (GeoJSON, KML) and visualization tools (Plotly, D3.js).
• Strong communication skills for technical-to-nontechnical translation and customer demos.
• Relevant cloud/AI certifications (AWS, Azure, Google, DeepLearning.AI).

Key Technologies & Platforms:

GPT 5, Claude, Gemini, Azure OpenAI Service, LangChain, LlamaIndex, Hugging Face Transformers, Python, Pinecone, ChromaDB, Weaviate, pgvector, AWS, Azure, Docker, Kubernetes, REST APIs, JSON/GeoJSON, Plotly, D3.js, IRIS, Maven, C2IE, OMEGA, FedRAMP IL2–4.

Target Salary Range: $135,000 - $216,000. This represents the typical salary range for this position. Salary is determined by various factors, including but not limited to, the scope and responsibilities of the position, the individual’s experience, education, knowledge, skills, and competencies, as well as geographic location and business and contract considerations. Depending on the position, employees may be eligible for overtime, shift differential, and a discretionary bonus in addition to base pay.

Benefits Statement: Peraton offers eligible employees a variety of benefits including medical, dental, vision, life, health savings account, short/long term disability, EAP, parental leave, 401(k), paid time off (PTO) for vacation, and company paid holidays. A full listing of available benefits can be viewed at https://www.careers.peraton.com/benefits.

Application Statements: The application period for the job is estimated to be 30 days from the job posting date. However, this timeline may be shortened or extended depending on business needs and the availability of qualified candidates. By applying to this job, you are expressing interest in the role and the Company. During the review of your application, you may be required to participate in an on-camera interview, as well as participate in a process to verify your identity. Use of artificial intelligence (AI) tools of any kind during Peraton interviews is strictly prohibited unless the candidate has obtained prior written authorization. All interview responses must be the candidate’s own.