TL;DR
Our client, a company building agentic AI products for enterprise customers, is hiring a Solution Architect, Agentic AI, in Bengaluru. You design agent workflows with customers, decide how agents connect to their tools and APIs, choose orchestration and guardrails, and take proofs of concept through to production. The role is hybrid, Bengaluru, with regular customer engagement.
The client is confidential at this stage. We will tell you who it is before your profile goes anywhere.
Key points
- Customer-facing architect for agent systems: workflow design, integration, orchestration and guardrails.
- Owns the path from proof of concept to production, not only the demo.
- Works alongside sales, forward deployed engineers and product, and translates between them.
- Technical depth is expected: you build prototypes and review production designs.
- Hybrid, Bengaluru. Confidential client, named to you before any submission.
What the work is
The Solution Architect, Agentic AI, turns a customer's business process into an agent system that works on their infrastructure and survives scrutiny from their security team. You lead the technical side of the engagement from first workshop to go-live.
The work covers five areas:
- Agent workflow design. Mapping a customer process, such as claims handling, procurement, reconciliation or support triage, and deciding which steps an agent should own, which need human approval, and where deterministic code is the better answer.
- Tool and API integration. Specifying how agents call the customer's systems: REST and GraphQL APIs, ERP and CRM platforms, document stores and internal services, increasingly exposed through Model Context Protocol servers. Authentication, rate limits and least-privilege permissions are part of the design, not an afterthought.
- Orchestration. Choosing between single-agent and multi-agent patterns and frameworks such as LangGraph, the OpenAI Agents SDK, CrewAI or Semantic Kernel, and designing state, memory, retries and handoffs so one failed step does not compound.
- Guardrails. Input and output filtering, prompt injection defences, action limits, audit logging and escalation rules that satisfy the customer's risk, compliance and InfoSec reviewers.
- Proof of concept to production. Scoping a proof of concept with honest success criteria, building or directing it, then hardening it with evaluation suites, observability, cost controls and a deployment plan the customer's IT team can operate.
What the client is looking for
An architect who has designed LLM or agent systems for real customers and can explain, in plain language, why an approach will or will not work. The strongest candidates can describe a proof of concept that never reached production and exactly why.
Must-haves
- Customer-facing solution architecture experience with enterprise clients.
- Hands-on LLM application work: tool calling, retrieval, structured outputs and evaluation.
- Integration depth across APIs, identity and access management, and enterprise data sources.
- Ability to build a working prototype yourself in Python or TypeScript.
- Clear written design documents and confident technical presentation.
Nice-to-haves
- Experience with one or more agent orchestration frameworks in production.
- Familiarity with cloud AI platforms such as AWS Bedrock, Azure OpenAI or Google Vertex AI.
- Domain knowledge in banking, insurance, logistics or another process-heavy sector.
Typical experience is usually 8 to 12 years, often from solution architecture, senior engineering or technical consulting, with recent hands-on LLM work.
How this role differs from adjacent roles
These titles are used inconsistently across the market. This is how the client separates them.
| Role | Main responsibility | Hands-on code | Relationship to customer |
|---|---|---|---|
| Solution Architect, Agentic AI (this role) | Design and technical lead of a customer's agent solution | Prototypes and reviews | Leads the technical relationship |
| Forward Deployed Engineer, Agentic AI | Builds and supports the deployment | Daily production code | Embedded with the customer |
| Enterprise AI Architect | Organisation-wide standards and platforms | Reference implementations | Internal stakeholders |
| Presales engineer | Technical win during sales | Demos | Pre-contract only |
| AI Engineer | Product features and pipelines | Daily | Rarely |
Process
The process mirrors the job: a real customer problem, designed in front of the people you would work with.
- A confidential conversation with SYTalent about the role and your background.
- A conversation with the client's head of solutions or engineering.
- A design exercise: architecting an agent workflow for a realistic enterprise scenario, including integrations, guardrails and a path to production.
- A customer-style presentation to the client's leadership, then references and offer.
Why this role is worth a look
Agent systems are moving from pilots to production, and the people who can design that transition for enterprise customers are hard to find. You will work across many industries and architectures, shape the client's reusable patterns from what you see in the field, and own outcomes customers can measure.
Related reading
Our AI, deep tech and quantum practice covers how these searches run, and multi-agent trading systems is a worked example of orchestration where errors are expensive. The Forward Deployed Engineer, Agentic AI and Enterprise AI Architect searches are the adjacent seats, and AI jobs in India maps the wider market.
Common questions
What does a Solution Architect for agentic AI do?
They design agent systems for specific customers. That means mapping the business workflow, deciding what agents should and should not do, specifying tool and API integrations, choosing an orchestration approach, setting guardrails, and leading the move from proof of concept to production. They are the customer's main technical counterpart through the engagement.
What skills do I need for an agentic AI Solution Architect role?
Enterprise solution architecture, hands-on LLM development including tool calling and retrieval, strong integration and security knowledge, and the ability to build prototypes yourself. Communication matters as much as technique: you will explain trade-offs to business owners, security reviewers and engineers, often in the same meeting.
How is this role different from a forward deployed engineer?
The Solution Architect leads the design and the technical relationship, and builds prototypes. The Forward Deployed Engineer writes and supports most of the production code inside the customer's environment. The two work closely together on the same accounts, but the architect spends more time on design decisions and the engineer more time on implementation.
Is the client named before I apply?
No, the client is confidential at this stage. If your profile fits, we will speak with you first and tell you who the company is before anything is submitted. Nothing moves without your agreement on this specific role. Because of application volume, only shortlisted candidates are contacted.
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.