Rajan Bhave

I spend most of my week with two kinds of people: customers trying to ship their first real GenAI workload on AWS, and AWS teams trying to figure out which partner to send those customers to. Somewhere in the middle of that Venn diagram is what I do at DoiT.

Concretely: I lead AI accelerators across EMEA (eight of them in the last year), for customers like Finlex, Emil, QueueIT, Belvo, Neteera, WealthWizards, Polarsteps and more. I am the architect who shows up at the kickoff, designs the Bedrock + AgentCore reference architecture, sits in the model-evaluation deep-dive, writes the production-readiness runbook, and then comes back six months later when the agent is in production and the bill needs explaining.

On the product side, I work shoulder-to-shoulder with DoiT engineering on the things I wish existed when I was in the customer seat: GenAI Intelligence (cost & usage across Bedrock, Vertex, OpenAI, Anthropic, Databricks and more), Snowflake Intelligence and the DoiT MCP Server on AWS Marketplace. On the partner side, I led DoiT's AWS Agentic AI Competency attainment, authored the Agentic AI Accelerator offer on AWS Marketplace, and drove the OpenAI → AWS GenAI Migration Accelerator to FTR + Marketplace approval — which is how DoiT landed in AWS's internal Book of Offers as "GenAI Migration Experts."

I talk about this work publicly — GenAI Loft Berlin, AWS Summit Zurich, AWS GenAI Roadshows in Munich / Zurich / Paris, the Social Dev Conference Hamburg, the Cloud Masters podcast with AWS — usually about the three things that decide whether a GenAI project survives contact with production: cost, identity, and autonomy.

If you're trying to figure out the difference between a GenAI demo and a GenAI product, that's the conversation I love having.