I’m Ivan, and I’m on the AI/ML Developer Advocate team at Google, focused on Vertex AI. My main goal here is simple: to be a resource for you. I’m here to help you find the answers, tools, and shortcuts you need to bring your ideas to life on Vertex AI.
What I enjoy most is getting my hands dirty with the messy parts of bringing research into the real world. I spend some time learning about how we can practically train, tune and serve open models (doing a lot with Ray lately) and how to build the necessary services to ship agents in production using the agentic stack (ADK, MCP, Agent Engine and A2A) on Vertex AI — things like prompt optimizers, memory bank, how to actually evaluate if these agents are any good and more.
On the side, I love working with startups , reading about cognitive neuroscience , and playing drums .
Anyway, I’m really here to learn from all of you. I’d love to hear about what you’re building and what challenges you’re running into. What are the interesting problems you’re trying to solve right now? And since Vertex AI is my day-to-day, please feel free to ping me directly with any questions. I’m always happy to help if I can.
It is great to connect with someone so close to the Vertex AI team. I genuinely love Vertex AI and have been using it every day since July 2025. I have essentially built my business with it, training and orchestrating agents entirely on Vertex AI alongside ChatGPT.
My biggest challenge now is token limits. I reached the limit right as my primary agent was close to being fully trained. I tried to continue in another session but it is not the same. I am close to hitting the limit again with a secondary agent and I am worried about losing progress. Is there a way to recover a previous chat or training state for my agents or to extend capacity in a structured way
Beyond this technical hurdle I would appreciate your perspective on how a startup like mine can deepen its relationship with Google. Should I explore a formal proposal to the Vertex AI team or is there a route for a small but ambitious venture to engage with Google for support whether through funding programs early partnership opportunities or registering on your partner platforms
Your guidance would mean a lot as I scale from a single developer operation into a structured venture.
I’m new to the Google Developer community and just saw your post – it really resonated with me. As an SME owner with a non-technical background, I’m looking to scale my business by developing my own products, and your focus on Vertex AI and deploying agents for practical applications sounds like exactly what I need.
I’m particularly interested in how Vertex AI can be used to optimize business operations through agent deployment. I’d love to connect and explore how we might collaborate on a project to bring this to life for my business.
I’m an old infra/sys/network/sec guy. My job is to deliver information systems to those who consume them. Obviously, the concept of “agent” is something that is particularly evident in its use when we have information systems whose variety has been evolving daily for 30 years… and especially whose levels of requirements have risen sharply as those who consume information systems have matured and observability allows clear visibility into what’s happening.
HI Ivan
I am aniket khorwal , I am from india
I am doing B.tech from National Institute of Technology , Silchar , Assam
I am passionate about machine learning , deep learning , MLops and software development
hope to learn so much from this community.
I’m Yeesian, and I’m on the Vertex AI Agent Engine team at Google; some of the issues that occupy my attention are at Vertex SDK and Google ADK (and beyond, e.g. A2A, AgentSpace).
Really appreciate Ivan setting up the relevant tags, and looking forward to collaborate with and learn from you all
I am Shawn (Xiaolong), I’m on the Vertex AI Agent Engine team at Google!
I focus more on the Serving side of Agent Engine (after you deploy your agent), and also Vertex SDK and Google ADK.
Great thanks to Ivan for the great talk and setting the forum up for our team.
Happy to work with you all!
, reading about cognitive neuroscience 
, and playing drums 
.
,