Data Scientist SME AI/ML

Leidos•Odenton, MD

15h

About The Position

Leidos Cyber & Information Sciences Division of the Cyber & Analytics Business Area is seeking an AI & Analytic Systems SME to serve as a Technical Closer for our COSS 3.0 program. You will architect the frameworks that achieve analytic superiority and actively perform the engineering required to push AI capabilities into production. You are expected to mentor senior technologists by example—working "fingers-on-keyboard" to solve the command's most complex technical roadblocks.

Requirements

Expertise in deploying LangGraph, CrewAI, or AutoGPT.
Hands-on fine-tuning of open-source models (Llama 3, Mistral) in air-gapped enclaves using PyTorch or TensorFlow.
Direct engineering of Retrieval-Augmented Generation (RAG) stacks using LangChain and vector databases (Milvus or Pinecone).
Proficiency in unified mission systems (e.g., Kudu Dynamic, TACMS, Maven Support Systems) with custom API integration skills.
Mastery of Python, Go, and Bash.
12–15+ years in technical roles.
5+ years specifically in Machine Learning Engineering or Data Science.
3–5 years working within the IC/DoD ecosystem, specifically with systems like Maven or TACMS.
Master’s Degree or PhD in Data Science, Artificial Intelligence, Computer Science, or Mathematics.
TS/SCI with CI Poly clearance required.

Nice To Haves

Google Professional ML Engineer, AWS Machine Learning Specialty, or NVIDIA Generative AI/LLM Associate certifications.
Military Equivalency: CMF Work Role Certification (Exploitation/Network Analyst) or experience as a 170A/17A.
A Bachelor’s degree with an additional 5+ years (totaling 15-20) of high-level engineering experience can often substitute for a Master's, especially if backed by the "Military Equivalency" noted in the description.

Responsibilities

Architect frameworks for analytic superiority.
Perform engineering to push AI capabilities into production.
Mentor senior technologists by example.
Solve complex technical roadblocks.
Deploy Agentic AI Frameworks (LangGraph, CrewAI, AutoGPT) for network defense.
Fine-tune open-source models (Llama 3, Mistral) in air-gapped enclaves using PyTorch or TensorFlow.
Engineer Retrieval-Augmented Generation (RAG) stacks using LangChain and vector databases (Milvus or Pinecone).
Integrate custom APIs with unified mission systems (Kudu Dynamic, TACMS, Maven Support Systems).