We are seeking an experienced AI Architect with 10+ years of technology experience to architect an Agentic AI led Project Setup Automation program for a global Life Sciences / CRO client.

This is a dual-track architect role : the candidate owns the traditional Azure application architecture — a React and Node workspace with its workflow, rules and action management services — and the AI architecture on top: deterministic defaulting, retrieval-grounded suggestion with citations, and a tool-using agent writing to Oracle Fusion PPM through Oracle Integration Cloud.

The ideal candidate will combine deep Azure application and integration architecture with hands-on LLM, RAG and agentic design, and a working functional understanding of Oracle Fusion and OIC .

Hands-on and client-facing, accountable for autonomy tiers, accuracy thresholds, the audit evidence record, and clearing client architecture and AI governance reviews under GxP and SOX.

Role and responsibilities

- Own the end-to-end solution architecture — Azure application services, the integration layer and the AI services — bound by one versioned canonical request schema, and architect the application stack: React and TypeScript SPA, Node / NestJS BFF with an OpenAPI contract, PostgreSQL, Redis, containerised on Azure App Service or AKS.
- Design the workflow and action management services — versioned state machine, gap-typed routing, SLA and tiered approvals, typed idempotent commands, and an immutable evidence record.
- Architect the AI services layer — deterministic defaulting via frequency tables with tie-break rules and reason codes, hybrid retrieval (dense plus BM25) with re-ranking and citations, and the evaluation rails:



golden set, accuracy thresholds, model registry and drift monitoring.
- Architect the agent runtime and planner with a tool registry over Oracle Integration Cloud REST endpoints: plan, confirm, execute, verify, write evidence.
- Define tiered autonomy and approval classes per tool — read auto-executes, low-risk on user confirmation, high-risk on named approval — with confidence banding and human-in-the-loop escalation.
- Apply functional understanding of Oracle Fusion PPM (project creation, templates, key members, DFFs, project-contract link) to field mapping, template selection rules and API coverage assessment.
- Work with the OIC team on orchestration flow contracts, mapping, fault handling, dead-letter queues and idempotent retries keyed on a workspace-issued correlation key.
- Own security, identity and observability — Entra ID SSO, OAuth 2.0 on-behalf-of with Fusion duty-role inheritance, secrets in Key Vault, OpenTelemetry tracing, and CI/CD with automated regression.
- Lead client architecture, security and AI governance reviews (ARB, AI Council, EIG), mentor the delivery POD, and stay hands-on through prototypes, the phase gate and handover.

Preferred Skills

- 10+ years in application and solution architecture, including 3+ years architecting AI or GenAI solutions.
- Proven delivery of both traditional enterprise applications and AI applications on Azure.
- Pharmaceutical, CRO, Healthcare or Life Sciences experience in GxP and SOX regulated settings.
- An agentic or RAG solution taken from proof of concept to production against measured accuracy.
- Microsoft Azure Solutions Architect Expert or Azure AI Engineer certification preferred.

📌 Technical Architect (Gurugram)
🏢 Incedo
📍 Gurugram