We are looking for a Chief Generative AI Engineer to join our team. As a Generative AI Engineer you will design, build, and operationalize a multi-agent AI contact center solution for a customer-facing conversational AI workstream. Scope covers LLM orchestration across specialized agents, prompt engineering and refinement, guardrail and compliance enforcement, automated evaluation and hallucination-detection pipelines, CI/CD evaluation gates, and production operations optimization for cost, latency, and containment. Delivery is fully AWS cloud-native, centered on Amazon Bedrock and Bedrock AgentCore with serverless orchestration (Lambda, Step Functions, EventBridge, DynamoDB, API Gateway) and integration into an Amazon Connect contact center flow.
Responsibilities
-
Architect multi-agent workflows on Amazon Bedrock AgentCore, defining agent roles, tool contracts, memory strategy, and hand-off patterns across the conversation lifecycle
-
Design prompt architectures, system instructions, and context-assembly patterns that hold up under production traffic and adversarial input
-
Define guardrail and compliance controls, including PII redaction, topic denial, grounding constraints, and safe-fallback paths, aligned to customer regulatory obligations
-
Build and iterate agent orchestration logic, tool/function integrations, and retrieval flows using Python on serverless AWS services
-
Implement structured prompt versioning and experimentation so that changes remain traceable and reversible
-
Integrate agentic capabilities with contact center telephony and chat channels, including deflection, escalation, and live-agent hand-off flows
-
Build automated test and evaluation pipelines using LLM-as-a-Judge, golden datasets, and rubric-based scoring for accuracy, tone, grounding, and task completion
-
Implement hallucination detection and grounding validation with quantified thresholds and regression tracking across model and prompt versions
-
Wire evaluation gates into CI/CD so that no prompt, model, or agent change ships without passing quality and safety criteria
-
Instrument observability, including traces, token accounting, latency, containment, and deflection rate, and drive tuning for cost and performance in production
-
Act as deputy to the Technical Lead, covering technical decision-making, customer-facing reviews, and delivery continuity during Lead absence, and mentor engineers on agentic and evaluation practices
Requirements
-
A minimum of 7 years of relevant experience
-
At least 2 years of leadership and team management experience
-
Experience in generative AI engineering, including LLM application design, prompt engineering, RAG, and model selection/tuning tradeoffs
-
Experience designing agentic and multi-agent workflows, including orchestration, tool use, memory, and hand-off patterns
-
Experience with LLM-as-a-Judge and automated evaluation, including rubric design, golden datasets, hallucination detection, and regression harnesses
-
Experience with Amazon Bedrock and Bedrock AgentCore
-
Strong Python skills for production AI services
-
Experience with AWS serverless architecture, including Lambda, Step Functions, API Gateway, DynamoDB, EventBridge, and CI/CD automation
-
Excellent English proficiency (B2 level or higher)
Nice to have
-
Experience with Amazon Connect, including contact flows, Lex integration, and Contact Lens
-
Familiarity with Responsible AI, guardrails, and compliance frameworks for regulated industries
-
Experience with observability and cost/latency optimization for LLM workloads
-
Experience with Infrastructure as Code tools such as CDK or Terraform
We offer
-
International projects with top brands
-
Work with global teams of highly skilled, diverse peers
-
Healthcare benefits
-
Employee financial programs
-
Paid time off and sick leave
-
Upskilling, reskilling and certification courses
-
Unlimited access to the LinkedIn Learning library and 22,000+ courses
-
Global career opportunities
-
Volunteer and community involvement opportunities
-
EPAM Employee Groups
-
Award-winning culture recognized by Glassdoor, Newsweek and LinkedIn
