Skip to content

Agentic engineering: Build and ship with Google Cloud

Read the Portuguese translation →

Author: Google Cloud · Watch on YouTube

Video authored by the channel shown, not by IANIA; all rights belong to the creator. Transcript and summary generated automatically; the original-language content is preserved.

Original language: English

Original video: Agentic engineering: Build and ship with Google Cloud · Channel: Google Cloud · 01:04 · Watch on YouTube

Topics: AI agents, AI evaluation and benchmarks, Software engineering with AI, AI infrastructure and hardware

The video introduces agentic engineering as a structured way to build and ship autonomous AI agents with Google Cloud. It compares the approach to supervised autonomous driving: agents handle execution while developers provide judgment, standards, and accountability. The course covers agent components and reasoning, a six-step lifecycle powered by CLI and ADK, and developer supervision through specs, evaluation, and monitoring. It concludes with a hands-on lab to build an agent and ship it to the Cloud.

  • Autonomous agents can query databases, enforce business policies, and launch campaigns.
  • Shipping agents to production is risky without structured engineering.
  • Agentic engineering keeps developers responsible for judgment, standards, and accountability.
  • The course introduces core agent components and the ReAct loop.
  • The agent lifecycle is described as a six-step process powered by CLI and ADK.
  • Developer supervision matters when authoring specs, evaluating quality, and monitoring telemetry.
  • The course ends with a hands-on lab to build an agent and ship it to the Cloud.
Start Chapter Summary
00:01 Why agentic engineering matters The video contrasts repetitive product-launch work with autonomous agents and explains why structured engineering is needed to reduce production risk.
00:32 The course journey Agentic engineering is likened to supervised autonomous driving, then the course introduces agent components, the ReAct loop, and a six-step lifecycle powered by CLI and ADK.
01:04 Supervision and shipping The course highlights authoring specs, evaluating quality, monitoring telemetry, and building and shipping an agent in a hands-on lab.

Because autonomy without structured engineering is a gamble. — 00:01

The AI agent handles the mechanical execution while you, the developer, stay firmly in the driver’s seat — 00:32

core agent components and how they reason through the React loop. — 00:32

Authoring specs. Evaluating quality and monitoring telemetry. — 01:04

Buckle up. Your journey starts now. — 01:04

agentic engineering, autonomous AI agent, ReAct loop, CLI, ADK, Google Cloud

00:01 Your team at symbol outdoor just spent 40 hours checking inventory, validating discount rules and writing promotional copy for one product launch again. Meanwhile, an autonomous AI agent can query your databases, enforce business policies, and launch campaigns in ECC. So why does shipping that agent to production still feel so risky. Because autonomy without structured engineering is a gamble. Welcome to agentic engineering with Google.

00:32 Think of agentic engineering as supervised autonomous driving. The AI agent handles the mechanical execution while you, the developer, stay firmly in the driver’s seat, supplying the judgment, standards and accountability. This course is your road trip. First, you get to know your vehicle. The shift to autonomy in the agentic era, core agent components and how they reason through the React loop. Next, you read the map the six step lifecycle

01:04 powered by CLI and ADK. Then you take the wheel at the three critical turns where your supervision matters most. Authoring specs. Evaluating quality and monitoring telemetry. Finally, you hit the road, building your agent and anti-gravity and shipping it to the Cloud in a hands on lab. Buckle up. Your journey starts now.