Turn AI pilots into secure, reliable production systems as a Solution Architect embedded with client business and engineering teams. You will design LLM, RAG, and agentic solutions, define governance and evaluation, and prove decisions with working code, helping enterprises adopt AI responsibly and at scale. Lead at EPAM, where your expertise guides global change. Collaborate with top talent, influence cutting-edge solutions, and create a legacy in a company that values leadership, innovation, and growth. Req.#1116590542 Responsibilities Partner with client business, product, engineering, security, and operations teams to resolve requirements and record architecture decisions Design end-to-end AI solution architectures covering RAG, agentic workflows and orchestration, Model Context Protocol (MCP) tool integration, Identity and Access Management, enterprise integration, LLM security, AI observability, and LLM evaluation Build Cloud FinOps cost models that link feature cost, usage, quality, and expected business value Document AI governance evidence for each agent, including autonomy limits, tool permissions, human checkpoints, evaluation and testing results, and rollback controls Model domain Knowledge Graph slices covering entities, relationships, business constraints, provenance, and owners, applying GraphRAG where relationships improve retrieval Build Python reference implementations and technical proofs with Forward Deployed Engineers, then preserve architecture integrity through the pilot-to-production transition under Architecture Governance and publish reusable patterns Requirements Client-facing solution architecture experience with business software or AI solutions that reached production, at either Solution Architect I level (own a bounded design with senior review) or Solution Architect II level (independently lead a complete client AI solution) Strong fundamentals in cloud computing architecture, microservice architecture, APIs, data flows, Identity and Access Management, and enterprise integration patterns Practical knowledge of LLM applications, Prompt Engineering, RAG, AI agent frameworks, tool calling through MCP, and agent orchestration Recent hands-on Python coding to prototype, inspect failures, and validate AI-assisted implementation work Experience defining LLM evaluation, AI observability, LLM security controls, and production quality attributes, plus an understanding of AI Governance, including agent permissions, autonomy, human oversight, auditability, and enterprise risk Clear architecture documentation and strong communication with business and technical stakeholders, with comfort working on-site amid ambiguity (regulated-industry and Cloud FinOps exposure is a plus)
