<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Gemini | Rabimba Karanjai</title><link>http://rabimba.fyi/tags/gemini/</link><atom:link href="http://rabimba.fyi/tags/gemini/index.xml" rel="self" type="application/rss+xml"/><description>Gemini</description><generator>Hugo Blox Builder (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Thu, 15 Oct 2026 14:00:00 +0000</lastBuildDate><image><url>http://rabimba.fyi/media/icon_hu_6d534efae7b021f2.png</url><title>Gemini</title><link>http://rabimba.fyi/tags/gemini/</link></image><item><title>LA Tech Week 2026 — Google for Developers Codelabs &amp; AI Workshops</title><link>http://rabimba.fyi/events/latechweek-2026-google-codelabs/</link><pubDate>Thu, 15 Oct 2026 14:00:00 +0000</pubDate><guid>http://rabimba.fyi/events/latechweek-2026-google-codelabs/</guid><description>&lt;h3 id="workshop-overview"&gt;Workshop Overview&lt;/h3&gt;
&lt;p&gt;At &lt;strong&gt;LA Tech Week 2026&lt;/strong&gt;, Google for Startups and Google for Developers hosted hands-on codelab tracks for enterprise founders, engineers, and technical leaders. Alongside fellow Google Developer Expert &lt;strong&gt;Kiruthika Subramani&lt;/strong&gt;, I co-instructed two interactive technical workshops detailing how to build production-grade agentic architectures with Google Cloud and the &lt;strong&gt;Model Context Protocol (MCP)&lt;/strong&gt;.&lt;/p&gt;
&lt;hr&gt;
&lt;h3 id="workshop-1-location-intelligence-adk-agent-with-mcp-servers"&gt;Workshop 1: Location Intelligence ADK Agent with MCP Servers&lt;/h3&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Interactive Presentation &amp;amp; Teleprompter&lt;/strong&gt;:
&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Official Google Codelab&lt;/strong&gt;:
&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Core Technologies&lt;/strong&gt;: Google Agent Development Kit (ADK), BigQuery Geospatial Data, Google Maps Grounding Lite MCP, Gemini 2.5 Flash&lt;/li&gt;
&lt;/ul&gt;
&lt;h4 id="architectural-focus"&gt;Architectural Focus&lt;/h4&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Agent Development Kit (ADK) Core&lt;/strong&gt;: Structuring autonomous agents with schema-enforced tool interfaces, state persistence, and error-recovery loops.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;BigQuery MCP Integration&lt;/strong&gt;: Bridging spatial data warehouses with natural language queries—enabling dynamic SQL generation, bounding box filtering, and spatial joins over enterprise logistics tables without manual query plumbing.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Google Maps Grounding Lite MCP&lt;/strong&gt;: Validating place details, computing multi-modal travel times, and grounding model answers against live physical coordinates.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Production Guardrails&lt;/strong&gt;: Rate-limiting spatial queries, handling disconnected MCP sessions, and sanitizing agent reasoning steps.&lt;/li&gt;
&lt;/ul&gt;
&lt;hr&gt;
&lt;h3 id="workshop-2-command--control-gemini-mcp--agentic-cli"&gt;Workshop 2: Command &amp;amp; Control: Gemini, MCP &amp;amp; Agentic CLI&lt;/h3&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Interactive Presentation &amp;amp; Teleprompter&lt;/strong&gt;:
&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Official Google Codelab&lt;/strong&gt;:
&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Core Technologies&lt;/strong&gt;: Antigravity CLI (&lt;code&gt;agy&lt;/code&gt;), Developer Knowledge MCP, Agent Skills, Nano Banana 2 Image Gen, Google Cloud Run&lt;/li&gt;
&lt;/ul&gt;
&lt;h4 id="architectural-focus-1"&gt;Architectural Focus&lt;/h4&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Headless Agent Orchestration&lt;/strong&gt;: Driving terminal-first agentic developer loops with Antigravity CLI (&lt;code&gt;agy&lt;/code&gt;), managing multi-step project scaffolding and debugging sessions.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Developer Knowledge MCP&lt;/strong&gt;: Exposing local codebase indexing, symbol trees, and architecture specs to agent runtimes without manual context dumping.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Multimodal Generation &amp;amp; Deployment&lt;/strong&gt;: Chaining multimodal synthesis via Nano Banana 2 for visual assets directly into automated container packaging and deployment to Google Cloud Run.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Composable Agent Skills&lt;/strong&gt;: Packaging reusable capability modules across teams to enforce testing, security, and deployment standards across agent prompts.&lt;/li&gt;
&lt;/ul&gt;
&lt;hr&gt;
&lt;h3 id="reception--industry-feedback"&gt;Reception &amp;amp; Industry Feedback&lt;/h3&gt;
&lt;blockquote class="border-l-4 border-neutral-300 dark:border-neutral-600 pl-4 italic text-neutral-600 dark:text-neutral-400 my-6"&gt;
&lt;p&gt;&lt;em&gt;&amp;ldquo;Google for Startups and Google for Developers take over #LATECHWEEK&amp;hellip; Thanks to my GDEs Rabimba Karanjai and Kiruthika Subramani for providing such wonderful and thought driven codelabs.&amp;rdquo;&lt;/em&gt;
— &lt;strong&gt;Marty Fortier&lt;/strong&gt;, Google Developer Ecosystem / Startup Lead&lt;/p&gt;
&lt;/blockquote&gt;
&lt;blockquote class="border-l-4 border-neutral-300 dark:border-neutral-600 pl-4 italic text-neutral-600 dark:text-neutral-400 my-6"&gt;
&lt;p&gt;&lt;em&gt;&amp;ldquo;LA TECH WEEK by a16z was hands-on. We built small tools around real questions from our day-to-day, like which truck a load needs and how to read proof-of-delivery paperwork faster, and got a map agent running in the cloud in a guided codelab&amp;hellip; Thank you to Rabimba Karanjai and Kiruthika Subramani, our instructors in the Google for Developers codelabs.&amp;rdquo;&lt;/em&gt;
— &lt;strong&gt;Elizabeth Moscoso&lt;/strong&gt;, Founder &amp;amp; Logistics Operator&lt;/p&gt;
&lt;/blockquote&gt;</description></item><item><title>Racecraft (Project Koru) · Prologue — The Origin Story</title><link>http://rabimba.fyi/blog/racecraft-project-koru-prologue-origin/</link><pubDate>Wed, 10 Jun 2026 00:00:00 +0000</pubDate><guid>http://rabimba.fyi/blog/racecraft-project-koru-prologue-origin/</guid><description>&lt;p&gt;I wrote about Project Koru (Racecraft) and how we built an AI Race Coach at Sonoma Raceway with Google Developer Experts. The post covers the origin story, the technical stack, and what we learned building real-time AI coaching at 100 mph.&lt;/p&gt;
&lt;p&gt;
&lt;/p&gt;
&lt;p&gt;Featured by Google:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;
— Google Developers Blog (July 2026)&lt;/li&gt;
&lt;li&gt;
— Google Developers Blog (January 2026)&lt;/li&gt;
&lt;li&gt;
— official Google video on YouTube&lt;/li&gt;
&lt;/ul&gt;</description></item><item><title>Google I/O GDE Field Test — AI Race Coach at Sonoma Raceway</title><link>http://rabimba.fyi/events/google-io-gde-field-test-sonoma-raceway-ai-race-coach/</link><pubDate>Sat, 23 May 2026 08:00:00 +0000</pubDate><guid>http://rabimba.fyi/events/google-io-gde-field-test-sonoma-raceway-ai-race-coach/</guid><description/></item><item><title>AI Race Coach (Project Koru)</title><link>http://rabimba.fyi/projects/ai-race-coach/</link><pubDate>Fri, 01 May 2026 00:00:00 +0000</pubDate><guid>http://rabimba.fyi/projects/ai-race-coach/</guid><description>&lt;p&gt;The AI Race Coach (Project Koru) is a real-time coaching system for motorsports, developed during the Google I/O GDE Field Test at Sonoma Raceway. The project demonstrated how agentic AI can bridge domain gaps between software engineering and racing expertise.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Architecture:&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Antigravity&lt;/strong&gt; — stateful orchestration and telemetry ingestion&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Gemini API&lt;/strong&gt; — cloud-based driver modeling and strategic insights&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Gemma 4 12B&lt;/strong&gt; — on-device edge intelligence for zero-latency audio coaching&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Jetpack Compose&lt;/strong&gt; — high-refresh Android dashboard for real-time visualization&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Google ADK&lt;/strong&gt; — multi-agent orchestration framework&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Custom USB Interface&lt;/strong&gt; — 10 Hz telemetry stream wired directly from vehicle sensors to Pixel 10&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;The system identified performance gains beyond human perception — including a new throttle application zone mid-corner in Turn 2 that secured a 0.1-second advantage. Featured on the Google Developers Blog and covered in my blog post &lt;em&gt;Racecraft: The Origin Story&lt;/em&gt;.&lt;/p&gt;</description></item></channel></rss>