AI Architect – Agentic AI Platform

THE OPPORTUNITY

Most companies claim to be building agentic AI. Very few are solving it at enterprise scale — where governance, auditability, and bounded autonomy are non-negotiable. Rivvun is building the platform that makes it real: autonomous AI agents that reason across business context, execute multi-step financial workflows, and operate within enterprise policy guardrails. Founded by former Icertis executive leaders, we are early, focused, and moving fast.

How do you architect bounded autonomy at enterprise scale? How do you design agent memory that is auditable and recoverable across long-running workflows? How do you build orchestration primitives that hold up when agents fail mid-execution? These are open questions. You define the answers.

We are looking for a AI Architect to own the technical foundation of this platform. Green-field. High ownership. Genuinely hard problems — with no existing architecture to inherit and no framework you are implementing for someone else.

WHAT YOU WILL WORK ON

• Design the agentic core. Build foundational platform capabilities: agent orchestration, lifecycle management, memory patterns, runtime controls, policy enforcement, and human-in-the-loop mechanisms.
• Define the reference architectures. Create the golden paths and reusable patterns for multi-agent workflows, bounded autonomy, and failure handling that the entire engineering organisation builds on.
• Build for enterprise trust. Architect secure-by-default AI systems with embedded governance, auditability, and access control that customers can stake their financial operations on.
• Create the platform layer. Design SDKs, workflow primitives, and shared services so teams spend time on problems, not re-implementing infrastructure. Drive architectural convergence and cut fragmentation.
• Stay at the frontier. Evaluate emerging agentic AI tools and frameworks continuously. Help Rivvun make sound, long-term technology choices — and know when to adopt versus when to build.

WHAT WE ARE LOOKING FOR

• 12+ years designing and operating complex cloud-native systems in production — with the scars to prove it
• Deep background in platform engineering, distributed systems, or enterprise architecture
• Hands-on experience building systems where LLMs or AI models are core components — not wrappers
• Strong track record of designing shared platforms and reusable frameworks that teams actually adopt
• Technical leadership through influence: design decisions, architecture reviews, cross-team alignment
• LLM system design: prompt orchestration, RAG pipelines, control flows, reasoning patterns, failure handling
• Agentic AI frameworks (LangChain, LangGraph, CrewAI, or equivalent) and a clear view on where they break at scale
• Multi-step AI workflow design: tool use, stateful memory, and long-running orchestration; vector DBs (Pinecone, FAISS, Weaviate)
• Cloud architecture (AWS, Azure, or GCP): security, networking, event-driven and async systems
• Security and governance: auth, authz, policy enforcement, and auditability for enterprise AI
• Strong Python; early-stage mindset — strategy and hands-on execution in the same week

WHAT SUCCESS LOOKS LIKE

• A durable, scalable agentic AI platform architecture — one that holds up as the product and customer base grow
• Reusable platform foundations that measurably accelerate engineering velocity across all AI initiatives
• Secure, governed, enterprise-ready AI workflows with technical clarity that the whole organisation builds on

WHY RIVVUN — WHEN YOU HAVE OTHER OPTIONS

• The problem is genuinely hard. Enterprise agentic AI has no textbook patterns yet. You will be working at the frontier — defining architectures that do not exist as standard references. That kind of work builds careers.
• Real ownership. No legacy system to maintain. You define the foundation — and the company builds on it.
• Founders who have done it before. Built by former Icertis executive leaders who scaled a category-defining enterprise software company globally. Not a first-time team figuring out enterprise commercialization.
• The window is short. Agentic AI is moving from experimentation to deployment. Getting in before the architecture is set means your decisions shape the product for years. Competitive salary, incentives, and early-stage equity.