Technology Architect

Overview

Architect who designs AI agent platforms for clients from start to finish, from understanding the client's business and current technology through to guiding delivery teams and building proofs of concept.

What you'll do

• Understand the client's business goals, current systems and pain points, and turn them into clear opportunities for agentic AI
• Follow changes in cloud agent services, AI models and agent standards (MCP, A2A), and judge which ones suit the client
• Identify quick wins that show value early, and plan a phased roadmap for the longer term
• Design the target architecture: agent orchestration, memory and state, tool and system integration, multi-tenancy, security, multi-region setup and disaster recovery
• Plan for users, request volume, response time, model usage limits and cost; decide how the platform scales and where it could fail
• Compare options (managed cloud agent services, open-source frameworks or custom build) and explain the trade-offs in simple terms
• Define the DevOps approach: environments, pipelines, release process, monitoring
• Set standards for agent security, guardrails, quality testing and cost control
• Support proposals, estimates and architecture reviews
• Manage client stakeholders: run workshops, present designs, handle concerns, align business and technology teams
• Build proofs of concept and reference implementations of agent platforms using agent frameworks and a cloud agent service
• Build agents for custom/legacy technology such as Java and Mainframe, and platforms such as SAP, Oracle and Workday
• Guide several delivery teams and check that what is built matches the design

What you'll need

• Minimum 15 years of experience
• Must have skill: Automation in Application Development
• 15 years full time education
• Hands-on experience building and deploying agents with at least one framework (LangGraph, Microsoft Agent Framework, Strands, Google ADK or similar)
• Hands-on experience running agents on at least one managed cloud agent service (AWS Bedrock AgentCore, Azure Foundry Agent Service or Google Gemini Enterprise Agent Platform)
• Deep in one cloud, with good knowledge of another: AWS Bedrock AgentCore, EKS, API Gateway, SQS, Step Functions
• Agentic IDEs: GitHub Copilot (GHCP), AWS Kiro, Claude Code
• Able to explain and defend design decisions to client CTOs and architects
• Comfortable leading across teams without direct authority
• Connects technical choices to business results
• Can show working agent solutions they have built, not only designs

Details

• Location: Bangalore office