Project Role: Solution Architect
Project Role Description: Translate client requirements into differentiated, deliverable solutions using in-depth knowledge of a technology, function, or platform. Collaborate with the Sales Pursuit and Delivery Teams to develop a winnable and deliverable solution that underpins the client value proposition and business case.
Must have skills : AI Agents & Workflow Integration
Good to have skills : NA
Minimum 15 year(s) of experience is required
Educational Qualification: 15 years full time education
Role Overview : -

We are seeking an experienced AI Agentic Architect to lead the design and implementation of enterprise-scale Artificial Intelligence, Generative AI, and Agentic AI solutions. This is a hands-on, player-coach role: you will split your time between architecting production-grade multi-agent systems and writing, reviewing, and setting the standard for the code and patterns your delivery teams build on.

Candidates must be genuinely code-fluent from day one — able to build and debug agentic systems personally, not direct from a distance. The ideal candidate will bridge business strategy and technical execution by defining scalable AI architectures, establishing governance standards, and driving AI adoption across the organization and its clients.

Roles & Responsibilities: Enterprise AI Architecture & Strategy

Define enterprise AI architecture and roadmap aligned with business objectives.
Design scalable and secure AI, GenAI, and Agentic AI solutions.
Design end-to-end AI solutions including RAG, AI Agents, Multi-Agent Systems, Knowledge Bases, and AI Workflows.
Architect AI integration with enterprise applications, databases, and cloud platforms.
Lead technical design reviews and architecture governance activities.

Ensure scalability, reliability, observability, and maintainability of AI solutions.

Agentic System Design (Hands-On)

Design Agentic AI ecosystems using orchestration frameworks, personally prototyping and hardening reference implementations.
Agent core mechanics: architect agent orchestration covering planning, reasoning, memory management, context management, task decomposition, and autonomous decision-making.
Multi-agent orchestration: design multi-agent patterns such as multi-tier topologies (router domain specialist utility agents), ReAct / ReWOO, and state-machine-based coordination (e.g., Lang Graph).
Tool-use & integration contracts: design tool/function-calling architectures and Model Context Protocol (MCP) integration layers, enforcing clean, decoupled contracts between agents, MCP servers, and downstream APIs.
Agent interoperability: define agent-to-agent communication and delegation using open interoperability standards (e.g., A2A)



so specialist agents built on different frameworks can discover and collaborate across accounts.
Architect AI copilots, autonomous agents, and workflow automation solutions.

Define LLM selection, evaluation, prompt strategy, and guardrails, including build/buy trade-offs across prompting, fine-tuning, and model adaptation for cost and accuracy.

Knowledge, Retrieval & Data Architecture
Implement enterprise knowledge management and semantic search capabilities.
Define data architecture, vector database strategy, and metadata management.
Advanced retrieval: architect retrieval strategies beyond baseline RAG — including Graph RAG, hybrid search, and knowledge-graph-grounded systems (e.g., Neo4j, Neptune, RDF/SPARQL) — for higher factual grounding.

Establish data governance, security, privacy, and compliance standards.

Evaluation, Observability & Reliability

Agentic evaluation: design evaluation pipelines that assess full agent trajectories — tool-choice correctness, argument validity, step count, grounding validation, hallucination detection, cost/latency, and policy compliance — not just final outputs.

Observability & tracing: define end-to-end tracing and runtime monitoring of agent decisions, tool calls, and outcomes, with continuous quality, cost, and success-rate metrics (e.g., Open Telemetry, Lang Smith / Langfuse, Datadog).
Reliability engineering: ensure resilience of agentic systems in production through deterministic state management, error handling, retries/idempotency, fallback and human-in-the-loop escalation, and budget/timeout controls.

Define mechanisms for model evaluation, risk management, explainability, and compliance.

Governance, Responsible AI & Cost
Establish Responsible AI principles and governance controls.
Agent lifecycle governance: define registration, configuration, versioning, deprecation, and runtime governance for agents, applying policy-as-code and identity-as-code across design-time and runtime.
Agentic cost governance (Fin Ops): design token-cost budgeting, blast-radius controls, and approval guardrails before autonomous actions with financial or operational impact.
Agent security: design defences specific to autonomous systems — prompt-injection mitigation, least-privilege tool-permission scoping, execution sandboxing, and input/output validation across the agent–tool boundary.





Ensure AI solutions meet security, legal, and regulatory requirements.

Leadership, Delivery & Reusable IP
Act as trusted advisor to clients and business leaders.
Lead architecture workshops and technical discussions.
Mentor AI engineers, solution architects, and development teams — leading by example on code quality and architectural standards.

Reusable solution IP: produce reference architectures, solution blueprints, integration guides, and sizing frameworks consumable by field delivery teams across concurrent client accounts.

Professional & Technical Skills: IT experience in AI/GenAI Architecture and hands-on building and deploying agentic AI systems in production.

Strong expertise in Generative AI, LLMs, RAG, AI Agents, and Agentic AI.
Agentic frameworks: hands-on experience with orchestration frameworks such as Lang Chain, Lang Graph, CrewAI, Auto Gen, Semantic Kernel, or equivalent, and agent design patterns (tool-calling, planning/reasoning, memory).
Hands-on experience with Azure AI, Azure OpenAI, Azure AI Foundry, AWS AI/ML, or Google Vertex AI.
Strong understanding of Machine Learning, NLP, Vector Search, and Knowledge Retrieval.
Experience designing enterprise-scale cloud-native architectures.
Expertise in Python, APIs, Microservices, Containers, and Kubernetes.

Agent operations & security: familiarity with agent evaluation, observability/tracing, and guardrail tooling (e.g., Lang Smith, Langfuse, Open Telemetry, NeMo Guardrails), and with agent security concerns such as prompt injection and tool-permission scoping.

Knowledge of security architecture, identity management, and AI governance.
Excellent communication and stakeholder management skills.
AI Architecture certifications from Azure, AWS, Google, or equivalent.

Knowledge of industry-specific AI use cases (Manufacturing, Aerospace, Automotive, BFSI, etc.).

Additional Information:
The candidate should have minimum 10+ years of IT experience, with 3+ years in AI/GenAI Architecture and 2+ years hands-on building and deploying agentic AI systems in production.
Experience in designing AI-powered enterprise solutions and digital transformation initiatives.
Demonstrated experience building and deploying agentic AI systems in production environments.
Ability to work with global and cross-functional teams.
Experience supporting proposal development, solution estimation, and client presentations.
Robust focus on Responsible AI, security, compliance, and governance.
Certification in AI Architecture, Cloud Architecture, or Enterprise Architecture is highly desirable.
This position is based at our Bengaluru office.

A 15 years full time education is required.

📌 Solution Architect (Bengaluru)
🏢 Accenture
📍 Bengaluru