Accountable for the end-to-end architecture, engineering blueprint, deployment model,
operational readiness, security, governance and integration strategy of client’s enterprise AI
systems — ensuring that AI solutions operate as secure, scalable, compliant and business-
aligned systems across models, applications, infrastructure, data, networks and external
dependencies.
That means the architect will have visibility across the entire chain:
Business → Business Rules → AI Platform → Models → Data/Knowledge grpahs →
APIs/Integration → Network → Infrastructure → Security → Deployment → Operations →
Monitoring
Key responsibilities
1. End-to-end AI system architecture
Define the overall architecture of AI systems across models, agents, applications,
data, APIs, infrastructure and external services.
Establish architecture principles, reference architectures and technology standards.
Define the technology selection principles and coding standards
Define the architecture standards to be followed by Individual AI products
Establish
Ensure architecture supports scalability, resilience, performance and maintainability.
2. Infrastructure & deployment
Define/ Approve the deployment architecture across cloud/ on-premise/ hybrid
environments both for Maveric and client environments
Define/ Approve requirements for Kubernetes, GPU infrastructure, storage,
networking and compute.
Establish/ Approve deployment, release and rollback patterns for AI systems.
Define checklists and guidelines to ensure production-readiness of AI products.
3. AI engineering ecosystem development
Define the standardized and reusable capabilities such as the following that needs to
be consumed by all Maveric AI products in a standardized manner
o Foundation models
o AI gateways
o Model routing
o Vector databases
o Prompt management
o RAG
o Guardrails
o Observability
o AI security
Define how Maveric AI platforms consume centralized platform capabilities
Work with Delivery and Inte