AI Solutions Architect

** Job Family:
** Data Science & Analysis
*
* Travel Required:

** Up to 10%
*
* Clearance Required:

** Active Top Secret (TS)
Guidehouse is seeking an AI Solutions Architect / Technical Lead to join our Technology / AI and Data team, supporting mission-critical initiatives for Defense and Security clients. In this role, you will serve as the principal architect for advanced AI platforms, driving end-to-end design across ingestion, retrieval-augmented generation (RAG), large language model (LLM) inference, and secure, scalable workflows. You will provide technical leadership and architectural governance to ensure coherence, performance, and compliance with stringent federal standards, while collaborating with engineers, data specialists, and mission stakeholders to deliver innovative solutions that enable data-driven decision-making.

This is an opportunity to shape cutting-edge AI capabilities that directly support national security objectives.
*
* What You Will Do:

*** Serves as the principal technical architect for the FBI adjudication AI platform, responsible for end-to-end architecture spanning ingestion, RAG, LLM inference, structured scoring, continuous learning loops, and analyst-facing mission workflows.
* Leads a dual-path LLM strategy that supports self-hosted open‑weight LLMs in AWS Gov Cloud as well as FedRAMP High managed model endpoints, ensuring secure, scalable, compliant, and mission‑aligned model operations.
* Provides architectural governance across backend services, model pipelines, data engineering, Dev Sec Ops , and UI/UX components to ensure coherence, performance, and SEAD‑4 adjudication alignment.
* Ensures platform architecture conforms to FedRAMP High, RMF, NIST 800‑53, and FBI ATO requirements, including boundary definition, logging standards, and continuous monitoring readiness.
** Architecture & Platform Leadership
*** Define the complete technical architecture for document ingestion, long‑document processing, embedding generation, vector search, anomaly detection, LLM inference, and adjudication memo generation components.
* Architect scalable infrastructure for 200M+ tokens/month processing volumes, supporting GPU‑based inference clusters, model routers, caching layers, and workflow orchestration services.
* Establish architectural standards for both self‑hosted GPU clusters and managed model endpoints, defining patterns for batching, KV‑cache utilization, parallel inference, and latency optimization.
* Produce architectural diagrams, sequence flows, integration contracts, and service interaction models to guide all development teams and maintain architectural consistency.
** AI/ML Workflow & Learning Strategy
*** Oversee design of AI/ML workflows supporting entity extraction, evidence linkage, SEAD‑4 scoring, anomaly detection, summarization, and document reasoning.
* Define fine‑tuning strategies using SFT, LoRA/QLoRA, dataset curation, evaluation pipelines, drift monitoring, and regression testing for consistent model quality.
* Establish standards for retrieval‑augmented generation (RAG), including embedding selection, retrieval indexing, chunking strategies, context assembly, and chain‑of‑thought management.
* Build monitoring patterns to detect hallucinations, scoring inconsistencies, bias, and model performance degradation across adjudicative domains.
** Security, Compliance & ATO Alignment
*** Ensure architecture supports FedRAMP High and RMF controls across data governance, encryption, access management, logging, segmentation, and monitoring.
* Author ATO documentation including SSPs, boundary diagrams, data flow diagrams, control implementation statements, and continuous monitoring artifacts.
* Embed secure‑by‑design engineering practices across APIs, data pipelines, LLM workflows, inference gateways, and audit logging layers.
* Define end‑to‑end auditability standards capturing prompts, model outputs, retrieval contexts, scoring rationales, and analyst overrides.
** Technical Leadership & Stakeholder Engagement
*** Provide senior engineering guidance to FBI leadership, translating complex model architecture, system behavior, and design tradeoffs into mission‑relevant insights.
*…