AWS Agentic AI Architect
Maarut
- Location
- Onsite (Toronto, Ontario)
- Employment
- Contract
- Level
- Senior Level
Posted 5 days ago
About the role
Maarut is hiring an AWS Agentic AI Architect to define reference architectures and governance standards for generative and agentic AI systems on AWS for a leading US life insurance and financial services client. The role involves leading architecture reviews, mentoring engineering teams, and scaling AI capabilities grounded in governed enterprise data.
Skills
AWS
Amazon Bedrock
SageMaker
Agentic AI
Generative AI
RAG
LLMOps
MLOps
LangGraph
Vector Databases
System Architecture
Cloud Architecture
Data Governance
Python
Serverless Computing
API Integration
Full job details
Job Description:
- We are hiring on behalf of a leading System Integrator
servicing a US life insurance and financial services company. The end-client is
over a century old, holds top-tier financial strength ratings, and serves
millions of families through life insurance, annuities, wealth management and
group benefits. It has rebuilt its data foundation on AWS, already runs
generative AI in production for its service teams, and is now scaling agentic
AI and modern engineering across the business.
- The client runs its data platform on AWS and already has
generative AI in production. It is now building a dedicated agentic AI
capability. As architect, you'll define how agentic and generative AI systems
are designed, secured, governed and scaled on AWS across the enterprise.
Key Responsibilities:
- Define the
reference architecture for agentic and generative AI on AWS: Amazon Bedrock
(models, Agents, Knowledge Bases, Guardrails), SageMaker, Lambda, Step
Functions, OpenSearch and DynamoDB.
- Architect
multi-agent systems: orchestration (LangGraph, Bedrock Agents or similar), tool
and API integration, MCP and agent-to-agent patterns, memory, and
human-in-the-loop controls.
- Design RAG
at enterprise scale: document pipelines, embeddings, vector databases, hybrid
retrieval, access-aware search and evaluation.
- Set LLMOps
and MLOps standards: model selection, evaluation, prompt and model versioning,
CI/CD, observability, cost management and incident response.
- Embed
security, privacy and responsible AI by design: data classification, PII
handling, guardrails, audit, and alignment with insurance regulation and model
risk management.
- Work with
data platform teams on Redshift, Glue, S3 and Lake Formation so AI is grounded
in governed enterprise data.
- Advise
senior stakeholders, lead architecture reviews and mentor engineering teams.
Requirements
Required
Qualifications & Technical Skills:
- 15+ years
in software, data or cloud architecture, with recent hands-on architecture of
generative and agentic AI systems in production.
- Deep AWS
expertise, especially Bedrock, SageMaker and serverless and data services.
- Strong
knowledge of LLM platforms, RAG, vector databases, agent frameworks and ML
fundamentals.
- Experience
with enterprise databases and data platforms, and with MLOps and LLMOps
tooling.
- A track
record of setting architecture standards and influencing executives.
Preferred
Qualifications:
- Experience
in insurance, banking or another regulated industry.
- AWS
Solutions Architect – Professional, or AWS Machine Learning / Generative AI
certification.
- Familiarity
with AI regulation and model risk frameworks for financial services.
Why Join
Us?
- Define how
agentic AI is built at one of the most respected names in US financial
services.
- Work on a
modern AWS foundation with real production AI already in place.
- Join at the
start of a strategic, executive-sponsored AI programme.