Experience:

• 8+ years in software engineering, platform engineering, or solution architecture
• 2+ years designing or building AI/LLM-enabled systems
• Production backend systems experience
• Enterprise environment experience with security, compliance, and change-management expectations
• Financial services experience preferred
• Distributed systems design experience
• Workflow or agent execution architecture experience
• Ability to translate ambiguous objectives into implementable architecture and delivery plans
• Advanced Python engineering skills
• Strong experience building API-first services
• Strong schema and data contract practice
• Deep understanding of reliable LLM interaction patterns
• Experience implementing and integrating MCP servers
• Experience with multi-step planning, routing, guardrails, and human-in-the-loop approvals
• Experience with grounding and retrieval patterns
• Strong understanding of different agentic frameworks
• Experience with test harnesses, golden datasets, and regression testing for prompts or agents
• Experience with safety testing, hallucination mitigation, and cost/performance controls
• Observability experience across traces, metrics, decision logs, prompt lineage, dashboards, prediction capabilities, and run replay

Responsibilities:

• Co-architect end-to-end agentic solutions
• Define engineering standards and best practices for safe, observable, testable agentic workflows
• Provide hands-on technical leadership through design reviews and reference implementations
• Support complex debugging and performance or scalability tuning
• Ensure enterprise readiness across DEV, UAT, and PROD
• Define solution patterns for routing, planning, tool execution, retrieval, approvals, and escalation
• Define multi-agent or multi-step workflows with deterministic control points
• Define evidence capture, explainability notes, and audit-ready outputs
• Create and enforce standards for prompt and agent versioning, structured output contracts, and validation
• Define safe tool execution boundaries with permissions and allow-lists
• Handle error handling, retries, idempotency, and compensating actions
• Define logging and tracing conventions and runbook expectations
• Define definition of done for production-grade agentic workflows
• Lead design reviews and support implementation teams
• Guide reference implementations, difficult integrations, performance bottlenecks, and incident triage
• Mentor Agentic Developers and uplift code quality through reviews and coaching
• Partner with security and client governance stakeholders
• Implement secrets handling, RBAC, identity integration patterns, and audit logging
• Define data residency and data handling controls, including redaction and masking
• Define change and release controls and environment promotion practices
• Ensure designs are compliant with regulated client expectations
• Define evaluation strategy for offline and online use
• Implement or guide observability practices
• Drive post-go-live optimization to reduce failures and improve determinism, latency, cost, and maintainability

Additional responsibilities:

• Co-own the reference architecture for agentic solutions, including components, interfaces, execution model, and security controls
• Collaborate with infra teams on on-prem and hybrid environments
• Work with developers, infra, and analysts on implementation support
• Maintain production-grade security, tests, observability, and documentation
• Support compliance and governance needs for BFSI-regulated clients

Nice to have:

• Experience with agent orchestration frameworks such as LangGraph-like patterns
• Experience with LLM observability tools such as Langfuse-like capabilities
• Experience working with open-source frameworks and customizing agent frameworks
• Experience with enterprise-hosted LLMs
• Experience designing vendor-agnostic model abstraction layers
• Exposure to orchestration tools such as n8n

More skills:

Platform architecture, LLM engineering, Agentic systems, Orchestration patterns, Service boundaries, Integration architecture, Non-functional requirements, Tool calling, Structured outputs, Approvals, Traceability, Security-by-design, Audit evidence, Data controls, Operational stability, Distributed systems, Service decomposition, API contracts, Asynchronous execution, Retries, Idempotency, Failure isolation, Resilience, Worker patterns, Job queues, Scheduling, State management, Execution traceability, Clean architecture, Modularity, Testability, Performance profiling, Packaging, Secure coding, Versioning, Backwards compatibility, Function calling, Safe tool execution, MCP servers, Multi-step planning, Routing, Guardrails, Human-in-the-loop approvals, Citation generation, Evidence generation, Prompt version management, Agentic frameworks, Test harnesses, Golden datasets, Regression testing, Safety testing, Hallucination mitigation, Cost controls, Performance controls, Observability, Application event tracking, Traces, Metrics, Decision logs, Prompt lineage, Dashboards, Prediction capabilities, Run replay, Logging, Monitoring, Production readiness, Capacity planning, Secrets management, RBAC, Audit logging, Environment separation, On-prem infra, Cloud infra, Hybrid environments, LangGraph-like patterns, LLM observability tools, Open-source frameworks, Enterprise-hosted LLMs, Vendor-agnostic model abstraction layers, n8n

Other:

• Job category: Agentic AI Architect
• Job type: Work from Office
• Location: Pune
• On-site/hybrid work mode
• Strong preference for on-prem, cloud, and hybrid infrastructure awareness
• Containers, Kubernetes/OpenShift fundamentals, logging and monitoring patterns, production readiness, capacity planning, secrets management, RBAC, audit logging, and environment separation are emphasized
• Enterprise security, governance, and risk controls are central to the role
• The role emphasizes regulated client expectations, especially BFSI
• Hands-on guidance for implementation teams and incident triage is expected