msampathkumar - Overview

Hi, I'm Sampath 👋

I build real-world demos, example apps and step-by-step guides for developers.

Most of my work focuses on improving the global developer experience for Gemini on Vertex AI. I embrace the philosophy of being a "First Customer, Last Engineer," shipping high-impact technical initiatives, code samples, documentation, and demos that help developers build AI applications seamlessly.

I maintain the Google Cloud - Gemini Cookbook Series, a step-by-step guide to teach you how to build Gemini features using real-world example apps like chatbots.

What I work on

AI Agents & LLMs: Building Agentic systems, multi-agent orchestration, and tooling using Google ADK and Vertex AI.
Developer Workflows: Establishing robust testing frameworks and evaluation (Evals) techniques for agentic reasoning and tool-calling accuracy.
RAG & Search: Deep diving into native multimodal embedding models (Gemini Embedding 2) and semantic search architectures.
Cost & Latency Optimization: Implementing Vertex AI context caching to make AI applications faster and significantly more cost-effective.
Open Source: Developing Python automations and data visualization tools, and maintaining repositories for developer resources.

About Me

Senior Developer Relations Engineer | Tech Lead at Google Cloud, specializing in enhancing the developer experience for Vertex AI Gemini.
Seasoned Lead Software Engineer with over 12 years of experience as a Python Big Data Engineer, handling Solution Architecture, Machine Learning, and Microservices.
Author & Technical Writer: I regularly publish on the Google Cloud Blog (e.g., Save costs and decrease latency while using Gemini with Vertex AI context caching) and maintain my personal technical blogs. I was also a Technical Reviewer for Hands-on Data Analysis and Visualization with Pandas.
Tech Speaker & Trainer: Active open-source contributor, technology trainer, and regular speaker at events like GDG CloudFest Poland.

🌱 Building agentic systems or looking to optimize your Gemini implementations? I'd love to connect and share insights!

Active Community Group: Linkedin