Job Requirements

- 10+ years in software engineering / architecture, with 3+ years designing AI-assisted or agentic software delivery at enterprise scale.
- Define the enterprise reference architecture for agent-orchestrated software delivery, shifting from SDLC to an agent-orchestrated development lifecycle.
- Design a layered agentic SDLC architecture spanning models/platform/infrastructure, engineering data and observability, foundation AI and knowledge, agentic control plane, SDLC agents and the engineering experience layer.
- Architect phase and generic SDLC agents across Requirements, Design, Coding, Code Review, Test, Release/SRE and Documentation, coordinated across CONTEXT → PLAN → BUILD → VERIFY → RELEASE → LEARN.
- Design the agentic control plane across orchestration, context/memory, task management, error recovery, human oversight, agent catalogue and A2A / event bus.
- Embed guardrails, agent sandboxing, prompt registry, skills hub, evaluation sets and model-risk controls into every stage.
- Define levels of autonomy (L1–L4) with human-in-the-loop / human-on-the-loop checkpoints and protected-execution boundaries.
- Integrate agents with enterprise SDLC toolchains including ALM/backlog, SCM and PR, CI/CD pipelines, test/ITSM and knowledge sources such as code/API repositories, standards, regulatory knowledge bases and test assets.
- Own cross-cutting trust, security, Responsible AI, IP provenance, data governance, audit and FinOps concerns across the lifecycle.
- Set standards, patterns and anti-patterns, and advise engineering leadership on the transition to AI-native software delivery.
- Define AgentOps, telemetry, tracing,



monitoring and observability frameworks for measuring agent performance, reliability, productivity impact and operational health.
- Architect enterprise knowledge systems leveraging Retrieval-Augmented Generation , vector search, context management and reusable knowledge services to improve agent effectiveness and accuracy.
- Establish agent lifecycle management practices including agent registration, versioning, deployment, certification, evaluation and retirement processes.
- Define enterprise AI platform strategies including model selection, model routing, cost optimization, capacity planning and AI service consumption governance.

Key Responsibilities

- Agentic SDLC Reference Architecture & Standards — define enterprise patterns, reusable architecture building blocks and migration guidance for AI-native software delivery.
- Agentic Control Plane & Multi-Agent Orchestration Design — architect orchestration, context, memory, task routing, recovery and agent-to-agent collaboration mechanisms.
- SDLC Toolchain & Knowledge Integration — integrate agents with ALM, source control, pull request, CI/CD, testing, ITSM and enterprise knowledge repositories.
- Autonomy, Guardrails, Evaluation & Human Oversight — define autonomy levels, approval gates, evaluation datasets, behavioural metrics and safe-execution boundaries.
- Trust, Security, Responsible AI & Governance across SDLC — embed security, IP provenance, Responsible AI, auditability, data governance and FinOps controls across the lifecycle.
- AgentOps, Telemetry & Observability - define monitoring, tracing, performance measurement, operational governance and continuous improvement frameworks for enterprise AI agent

📌 Agentic SDLC Architect (Mumbai)
🏢 Tata Consultancy Services
📍 Mumbai