TL;DR
Our client, a large technology-led business with several business units, is hiring an Enterprise AI Architect in Bengaluru. The role defines the reference architecture for LLM and agentic systems across the organisation: models and vendors, retrieval and data, security, governance, evaluation and cost. It is a senior individual contributor seat with wide influence. The role is hybrid, Bengaluru.
The client is confidential at this stage. We will tell you who it is before your profile goes anywhere.
Key points
- Owns the enterprise reference architecture for LLM, RAG and agentic systems across business units.
- Leads model and vendor selection, with cost, latency, data residency and lock-in weighed explicitly.
- Sets the standards for governance, security and evaluation that every AI team builds against.
- Senior hands-on architect: you prototype, review designs and unblock teams, not only write documents.
- Hybrid, Bengaluru. Confidential client, named to you before any submission.
What the work is
The Enterprise AI Architect stops each business unit from building its own incompatible AI stack. You define the shared patterns, platforms and guardrails so teams ship faster and the organisation carries less risk.
The responsibilities fall into six areas:
- Reference architecture. A documented, working blueprint for LLM applications and agentic systems: gateways, orchestration, retrieval, tool access, memory, observability and human approval steps. Teams should be able to start from it rather than from a blank page.
- Model and vendor selection. Comparing hosted models from OpenAI, Anthropic and Google with open-weight models such as Llama, Mistral or Qwen served on internal infrastructure, and platforms such as AWS Bedrock, Azure OpenAI and Google Vertex AI. You set the criteria and run the evaluations, rather than following whichever vendor called last.
- RAG and data architecture. How enterprise content is chunked, embedded, indexed and permissioned; when to use a vector database, hybrid search or a knowledge graph; and how retrieval respects existing access controls so an agent never surfaces a document its user could not open.
- Governance. Model inventories, approval paths, usage policies and documentation that stand up to internal audit and customer due diligence, aligned with the NIST AI Risk Management Framework, ISO/IEC 42001 and India's Digital Personal Data Protection Act.
- Security. Threat models for prompt injection, data exfiltration through tools, over-permissioned agents and supply chain risk in models and packages, drawing on the OWASP Top 10 for LLM Applications.
- Evaluation and cost. Common evaluation harnesses and quality gates before release, plus a cost model covering tokens, inference hardware, caching and model routing so finance can forecast AI spend.
What the client is looking for
A hands-on architect who has taken AI systems into production across more than one team, and who can persuade business unit leaders without formal authority over them. The strongest candidates can show an architecture decision they made, the alternatives they rejected, and what happened afterwards.
Must-haves
- Architecture ownership of production LLM or ML systems in an enterprise setting.
- Depth in retrieval-augmented generation, embeddings, vector search and data access control.
- Working knowledge of at least one major cloud AI platform and of serving open-weight models.
- Clear thinking on security and governance for AI, including agent permissions and tool access.
- Ability to write and present architecture to engineers and executives alike.
Nice-to-haves
- Experience designing multi-agent or tool-using systems with frameworks such as LangGraph or Semantic Kernel.
- Background in enterprise architecture frameworks such as TOGAF.
- Exposure to regulated sectors, for example banking, insurance or healthcare.
Typical experience is usually 12 to 16 years in software and data architecture, including recent years focused on LLM systems.
How this role differs from adjacent roles
The title overlaps with several others in the market, so this is how the client draws the lines.
| Role | Scope | Main output | Customer-facing |
|---|---|---|---|
| Enterprise AI Architect (this role) | Whole organisation, all business units | Reference architecture, standards, platform choices | Internal stakeholders |
| Solution Architect, Agentic AI | One customer or programme | A specific agent solution design | External customers |
| Head of Artificial Intelligence | AI strategy and team | Roadmap, team, business outcomes | Board and executives |
| Cloud or data architect | Infrastructure or data platform | Cloud and data designs | Internal |
| AI Engineer | Individual features | Production code | Rarely |
Process
A structured process that tests architecture judgement on real problems.
- A confidential conversation with SYTalent about the brief and your experience.
- A meeting with the client's technology leadership.
- An architecture case study: designing an agentic system for one of the client's business units, including model choice, retrieval, security and cost.
- Conversations with business unit stakeholders, then references and offer.
Why this role is worth a look
The seat sets direction for AI across an entire organisation at the point where decisions are still being made, not inherited. You will see a wide range of use cases, work across business and engineering leadership, and build the shared foundation other teams depend on. Architects who can do this well for agentic systems are scarce, and the experience compounds.
Related reading
Our AI, deep tech and quantum practice explains how these searches run, and agentic AI in trading shows agent design under strict reliability limits. The Head of Artificial Intelligence and Solution Architect, Agentic AI searches are the adjacent seats, and AI jobs in India maps the wider market.
Common questions
What does an Enterprise AI Architect do?
An Enterprise AI Architect defines how an organisation builds and runs AI systems. That includes the reference architecture for LLM and agentic applications, model and vendor selection, retrieval and data design, security and governance standards, evaluation gates and cost controls. The aim is that every business unit builds on shared, safe foundations instead of separate stacks.
How is an Enterprise AI Architect different from a Solution Architect?
An Enterprise AI Architect works across the whole organisation and sets the standards and platforms. A Solution Architect designs a specific solution for one customer or programme within those standards. The enterprise role is broader and more internal. The solution role is narrower, usually customer-facing, and closer to delivery of a single system.
What skills are needed to become an Enterprise AI Architect?
Strong software and data architecture, hands-on depth in retrieval-augmented generation and LLM systems, knowledge of cloud AI platforms and open-weight model serving, and a working grasp of AI security and governance. Equally important is the ability to explain trade-offs to executives and persuade teams you do not manage.
Is this role hands-on?
Yes. The client wants an architect who prototypes, reviews code and designs, and helps teams through difficult implementations, not one who only produces slide decks. Expect to build reference implementations and evaluation harnesses yourself, alongside setting standards and advising business unit leadership on their AI plans.
Interested in this role?
Apply once, to us. We confirm the client with you before anything is submitted. Because of application volume, only shortlisted candidates are contacted.