Job overview
Job Title: Agentic AI Architect
Location: Bangalore
Experience: 10–15 years (at least 3 years in GenAI / LLM / agentic AI)
Reporting to: Head of AI/ML
We are looking for an Agentic AI Architect to design and lead endtoend architectures for AI agent–based systems that can reason, plan, and act autonomously across complex business workflows. You will define the blueprint for multiagent ecosystems, select the right LLMs and orchestration frameworks, and work closely with product and engineering teams to take solutions from experiment to production.
About the role
In this role, you will own the overall architecture of our agentic AI platforms, from knowledge management and retrieval to agent orchestration, tools, and system integration. You will be the primary technical leader for how AI agents interact with enterprise data, APIs, and business processes, ensuring scalability, reliability, safety, and cost efficiency.
You will collaborate with AI/ML engineers, data engineers, solution architects, and business stakeholders to translate use cases into robust, productiongrade designs. The ideal candidate combines strong software/solution architecture experience with handson exposure to LLMs, RAG, and modern agentic frameworks.
Key responsibilities
- Define the architectural vision, principles, and reference patterns for agentic AI solutions across the company.
- Design endtoend architectures for agentic AI systems, including LLM integration, knowledge management (RAG, hybrid search), and tool/agent orchestration.
- Architect multiagent systems and autonomous workflows where AI agents can plan, decompose tasks, use tools, and coordinate to achieve business goals.
- Evaluate and select models (opensource and commercial), orchestration frameworks (e.g., LangGraph, CrewAI, AutoGen, Semantic Kernel, Vertex/Bedrock tools), vector DBs, and other core components.
- Design and oversee RAG pipelines, including document ingestion, chunking strategies, embeddings, indexing, and retrieval patterns tailored to agent use cases.
- Define safety, governance, and observability guardrails: prompt policies, cost and latency budgets, auditability, decision traceability, and fallback mechanisms for agent actions.
- Partner with engineering teams to integrate agentic AI services with existing applications, data platforms, APIs, identity, and security layers.
- Lead technical discovery and solutioning for new use cases, run PoCs/experiments, and drive successful handoff into scalable production implementations.
- Establish best practices, standards, and reusable blueprints for building, testing, and deploying agentic AI systems.
- Mentor engineers and architects on agentic AI concepts, architecture patterns, and implementation approaches.
Required qualifications
- 8–12 years of total experience in software/solution/AI architecture, with significant time designing distributed or cloudnative systems.
- 3+ years working with modern AI/ML systems, including handson exposure to LLM applications (chatbots, copilots, agents) in production or advanced pilots.
- Strong understanding of agentic AI concepts: autonomous agents, planning and tool use, multiagent coordination, and feedback loops.
- Experience designing systems using LLM orchestration frameworks (e.g., LangGraph, Semantic Kernel, AutoGen, CrewAI) or major cloud agent platforms (Azure AI Agent Service, Google Vertex agents, AWS agentic services).
- Solid knowledge of RAG architectures, vector databases, and hybrid search (BM25 + dense retrieval, reranking, etc.)
- Proficiency in at least one modern programming language used for AI applications (Python preferred; TypeScript/Java/Go as a plus) and in designing APIs and microservices.
- Experience with cloud platforms (AWS, Azure, GCP) and infrastructure patterns for AI workloads (containerization, serverless, GPUs, monitoring, observability).
- Strong architectural thinking: ability to balance tradeoffs between cost, performance, safety, and timetomarket.
- Excellent communication skills; able to explain complex AI concepts to both technical and nontechnical stakeholders and lead architecture discussions.
Preferred qualifications
- Prior experience architecting multiagent systems or complex workflow automation using agents acting over enterprise tools (CRM, ERP, ticketing, developer tools, etc.)
- Experience with prompt engineering, evaluation frameworks, and automated testing for LLM/agent systems (e.g., guardrail frameworks, eval harnesses, synthetic testing)
- Background in data engineering or MLOps/LLMOps, including CI/CD for AI, feature stores, and monitoring of model/agent performance and drift.
- Knowledge of Responsible AI practices and emerging AI regulations; experience embedding compliance and governance requirements into solution designs.
- Industry domain experience in [your focus verticals – e.g., financial services, healthcare, SaaS, industrial, etc.] to design agents aligned with realworld workflows.
What you will work on
- Design an agentic “copilot” that orchestrates multiple agents to research, plan, and execute complex internal processes endtoend (e.g., onboarding, incident response, sales operations).
- Architect a retrievalaugmented, multitool agent that reads from knowledge bases, interacts with internal APIs, and takes safe actions with approvals and guardrails.
- Build a reusable platform foundation for future AI agents, including shared libraries, policies, and observability so new use cases can be added quickly.
Equal opportunity statement
We are an equal opportunity employer and value diversity at all levels of the organization. All qualified applicants will receive consideration for employment without regard to gender, gender identity, sexual orientation, religion, disability, age, marital status, caste, or any other protected characteristic, in line with applicable laws in our operating
