Client-Facing AI Architect: Take AI to Production

EPAM Systems Inc; United States; Remote; USD 170,000 - 230,000; Full time

Job summary

EPAM Systems seeks a Forward Deployed AI Architect to embed with client business owners and engineering leaders, designing AI-native, enterprise AI and applied AI solutions that move from pilot to production. You will own the architecture, framing problems, and making trade-offs in the client’s vocabulary and timeline.

You will embed with client teams, define system scope, and determine what can be delegated to AI versus human oversight, including cost models, guardrails, and evaluation criteria.

Qualifications

• Experience taking AI-native systems to production at a client with real users.
• Proven client-facing architect who frames problems and presents cost–risk options.
• Solid experience in integration, distributed systems, cloud, and data-flow at enterprise scale.
• Willingness to be on-site and travel as required.

Responsibilities

• Embed with a client account, working with business owners and engineering teams to move AI from pilot to production.
• Design end-to-end AI-native architectures including orchestration, retrieval, and guardrails.
• Own artifacts to measure impact: capability economics, trust-rung evidence, and knowledge-graph slices.
• Turn divergent requirements into explicit business decisions with named owners.
• Define evaluation criteria to ensure business scope alignment, not just fluent outputs.
• Build proofs of concept and reference implementations when needed.
• Contribute learnings back to EPAM's AI architecture practice.

Skills

Enterprise AI architecture

Solutions architecture

RAG/agentic workflows

Cloud & data-flow

Client-facing consulting

Hands-on coding

Ambiguity tolerance

Tools

OAuth 2.1

SPIFFE

LLM observability

Job description

EPAM Systems seeks a Forward Deployed AI Architect to embed with client business owners and engineering leaders, designing AI-native, enterprise AI and applied AI solutions that move from pilot to production. You will own the architecture, framing problems, and making trade-offs in the client’s vocabulary and timeline.

You will embed with client teams, define system scope, and determine what can be delegated to AI versus human oversight, including cost models, guardrails, and evaluation criteria.