标签：#googlecloud

I created this content for the purposes of entering the All Things Agentic Hackathon. The problem that started it A doctor gets about eight minutes with a patient and, for anyone with a real history, forty pages of scattered records — lab reports, discharge notes, and pharmacy bills from three different clinics. So the history is effectively invisible at the exact moment it matters most. And when the visit ends, nothing follows up: the six-month course lapses at week five, the recheck never gets booked. I wanted to build an agent that closes that loop — one that reads the mess, decides urgency in a way a clinician can actually trust, and handles the follow-up on its own. That became CareLoop , my entry for the All Things Agentic Hackathon (Taskmaster track), built on Gemini, the Google Agent Development Kit (ADK), Cloud Run, and Firestore. The one principle I wouldn't compromise on Rules decide, AI explains. The temptation with an LLM is to let it do everything — including deciding whether a chest-pain patient is urgent. I refused to do that. In CareLoop, a deterministic engine owns every clinical decision: a weighted symptom score plus a red-flag override sets the triage level and routing. It is fully auditable, and it returns byte-identical output on the same input every single time. The LLM's job is strictly language: Reading unstructured documents into a fixed schema — I call it "Gemini extracts, rules merge." Writing the structured result into a plain-language brief a clinician can skim in ten seconds. No language model is ever in the decision path. When a judge asks "why was this Critical?", the answer is a score breakdown they can inspect — not a model's say-so. That single decision shaped the whole architecture. What it actually does CareLoop runs the full loop end to end: Ingest & compact — it reads a patient's documents and merges them into one structured ledger: allergies, chronic conditions, active medications, and lab trends over time. Instead of pushin

Every workflow builder I have used opens the same way: a blank canvas and a palette of nodes. Zapier, n8n, Make - all of them assume you already know what you want, already decomposed into steps, before the tool is any use to you. Most people don't. They know the chore . "I keep forgetting to check the weather before I bike in." The gap between knowing the chore and knowing the DAG is precisely the work these tools leave you to do alone, and I think it is why most people who try one never build a second automation. So I built Weaver, which inverts it. Weaver interviews you about the chore, one question at a time, until it actually understands the goal. Then it designs the workflow, validates it, deploys it, and runs it. The canvas is an output rather than an input. This post is about the parts that did not go to plan, because those are the parts worth reading. The interview is the whole product Three rules, and they are harder than they look: One question per turn. Never three bundled into a paragraph. Never invent a value the person has not given you. No quietly assumed recipient, city, or time. A correction updates one detail. Say "actually, Mondays" halfway through and it changes that and keeps going, instead of restarting the interview. That third one is the one people notice. Restarting an interview because the user corrected themselves is the single fastest way to make software feel like it is not listening. Only once it restates the whole task in plain language and you confirm does it save the intent and hand off to a separate Designer Agent. Two agents, deliberately not one The Conversation Agent and the Designer Agent are different models with different prompts and no shared state beyond a saved intent. That is a design decision, not an accident of implementation. Understanding a person and designing a system are different skills with different failure modes. Collapsing them into one prompt makes both worse: the interviewer starts proposing architecture hal

Upgrading Kubernetes clusters across a large enterprise fleet is often a balancing act between staying current with security patches and avoiding outages. By default, Google Kubernetes Engine (GKE) rolls out automatic upgrades progressively according to Google Cloud regional timelines. While regional rollout works well for standalone clusters, it does not understand your organization's business topology. If you run staging clusters in us-central1 and critical production clusters in us-east1 , a standard regional rollout could upgrade your production environment before your pre-production validation completes. The General Availability (GA) release of GKE rollout sequencing with custom stages solves this challenge. It provides platform teams with declarative control to sequence cluster upgrades across fleets, environments, and even distinct Google Cloud organizations according to business criticality rather than cloud geography. How rollout sequencing works Rollout sequencing builds on GKE fleet management. Fleets serve as logical boundaries for environments such as development, staging, and production. With rollout sequencing, you define an ordered pipeline of upgrade stages managed by a central resource called RolloutSequence . When GKE publishes a new automatic upgrade target for a release channel, or when you explicitly trigger a target version, the system creates a Rollout object. This rollout progresses through your defined stages sequentially: Control plane upgrades start in the first stage. Once all control planes in that stage reach the target version, a stage soak timer begins. Node upgrades run in parallel with control plane upgrades, respecting node pool upgrade strategies such as surge or blue-green. When both control planes and nodes complete their upgrade and satisfy the configured soak duration, the rollout advances to the next stage in the sequence. If an individual stage contains clusters that take longer than 30 days to finish upgrading—due to restr

Link to the Source Code As a data engineer, I spend my days designing clean, optimized data structures. But at home, I face a much tougher crowd: my family. We manage our shared finances together to optimize our budget, and because of where we live and work, we have to do this in several different currencies (like USD, EUR, CZK, and UAH) Like any developer, I first tried to find a ready-made app to solve this. But I ran into a classic problem: they were either bloated with a million features we didn’t care about, or they were missing the exact features we actually needed. So, we did what any desperate family does: we opened a Google Sheet . We tracked our money there for a while, not because it was perfect, but because it helped us figure out what we actually needed from a real application. It was our "living schema design" before I wrote a single line of code. In this article, I want to show you how I looked at this problem from two sides—as a frustrated user who just wants to log expenses, and as a data engineer obsessed with clean database design. Here is the story of how I built our custom home accounting server. Part 1: Django, a Star Schema, and the Framework Battle By 2025, I was ready to replace our Google Sheet. My main programming language is Python, so I had three realistic choices: FastAPI, Flask, or Django. FastAPI is the cool kid on the block for high-speed APIs, but we didn’t expect millions of requests (unless my family suddenly grew by a factor of a million). We also needed a friendly web UI, which FastAPI isn't naturally built for. I had just used Flask for my previous project, but I wanted to challenge myself and learn something new. Django felt like an old friend I hadn't seen in years. It has amazing built-in tools (like the admin panel and great translation support), and using it was the perfect way to refresh my skills and grow as a developer. The Database: Why a "Star Schema" Actually Makes Sense As a data engineer, I didn’t want a messy data

The correlation notice fired. Three sites, same anomaly type, inside the time window. The orchestrator caught it and logged it, live, against the deployed service. Clean, first try. Then I asked myself a question I almost didn't bother asking, because the thing had just worked: why did it work? The answer wasn't "because the logic is correct." It was "because Cloud Run happened to route both requests to the same running instance." Well, shit. My orchestrator was holding its list of recent risk events in a plain Python list, in process memory. Worked in local testing because there's only one process. Worked live because Cloud Run, under light traffic, often reuses the same instance instead of spinning up a second one. Neither one's a guarantee. The moment traffic patterns shifted and two readings landed on two different instances, the second instance wouldn't have a clue the first one existed. A correlation that should fire would just silently not. A bug that passes its own demo is the hardest kind to catch. There's no error to chase. There's just a checkmark. What I was building VES Fleet is a network of independent site-agents (Bori, Choba, Etche, three real survey sites in the Niger Delta). Each one reads an underground electrical survey, send current into the ground, measure how it flows back, a real physical signal of what's down there, and calibrates its own contamination-risk threshold from its own site's real history. Not a number copied from anywhere else. An orchestrator watches for the same risk signature showing up at more than one site inside a time window. It's my submission to the Fortified Enterprise Fleet track of Google's All Things Agentic Hackathon. Architectural discipline is 30% of the score there. Proving it actually runs on Google Cloud is a separate 30%. So a bug that only looked fixed was never going to survive someone actually reading the state-management story. Checking the thing that already worked Once I understood the actual failure mod

Gen AI Academy APAC — Track 1 (AI Agents with Gemini, ADK, and Cloud Run) Why I built this I lead a customer service team of 25 agents at Amazon, handling both buyer-side and marketplace seller support. A big part of my job is reviewing escalated cases — calls or chats where a customer asked for a supervisor — and figuring out why they escalated in the first place. Was it a policy gap? A training issue? A system limitation nobody flagged? Right now, that review is manual. Every escalation gets read, tagged, and turned into a coaching note by a human — usually me, or one of my leads. It works, but it doesn't scale well, and patterns across dozens of cases are easy to miss when you're reviewing them one at a time between everything else on your plate. So for Track 1 of the Gen AI Academy APAC program, I built an agent that does the first pass of this analysis automatically: read an escalation summary, classify the root cause against a standard taxonomy, flag whether it looks like a repeat pattern, and draft a coaching note — the same way I would, just faster and more consistently. What it does The agent takes a case summary like this: Customer requested a refund for a damaged item outside the standard return window. Agent denied it citing policy; customer says a rep last month approved a similar exception for someone else. And returns a structured analysis: { "root_cause_category" : "policy_misapplication" , "severity" : "medium" , "is_likely_repeat_pattern" : true , "pattern_reasoning" : "Inconsistent policy application across agents suggests a training or documentation gap rather than an isolated error." , "coaching_note" : "..." } It's built on Google's Agent Development Kit (ADK) with Gemini as the underlying model, and deployed as a live service on Cloud Run . The agent has one tool — a lookup function for the standard root-cause taxonomy — which keeps the categories consistent and easy to update without touching the core prompt. For batch review, I also built a

In our inaugural issue , we established that the future of enterprise AI lies not merely in raw model parameters, but in the architectural paradigms—specifically Graph Neural Networks (GNNs)—that capture relational intelligence. However, the most sophisticated architectural decision is rendered obsolete if the deployment infrastructure introduces prohibitive latency. At Informatiqs, we emphasize that model deployment is fundamentally an operations research problem. As we transition from batch-processed predictions to real-time Generative AI and dynamic Machine Learning on Google Cloud Platform (GCP), we confront the inherent friction between compute elasticity and system responsiveness: the notorious "Cold Start" problem. In this issue, we dissect the mathematics of serverless inference, the orchestration of Cloud Run and Eventarc, and how minimizing initialization latency is the ultimate enabler for high-frequency, event-driven enterprise intelligence. 1. The Mathematical Anatomy of the Cold Start To engineer a solution, we must first formalize the problem. In a serverless architecture (scale-to-zero), infrastructure scales dynamically with demand. The total response time for an inference request can be understood as a composite of three phases. First, the baseline network latency. Second, the actual inference time—the computational effort of the model itself. The critical variable, however, is the conditional penalty phase. If a serverless container has scaled to zero, the system must endure the time required to provision new compute resources and the heavily taxing process of loading massive neural network weights into memory. If the container is already 'warm', this penalty is completely bypassed. We can model the probability of encountering this cold start using queueing theory. Assuming incoming inference requests arrive as a stochastic process, the likelihood of a cold start is determined by the mathematical relationship between the frequency of incoming requ

In the current epoch of Artificial Intelligence, the industry remains singularly preoccupied with the "Model" — obsessing over the raw parameter scales of the latest LLMs or the specific benchmark performance of a new transformer variant. However, at Informatiqs, we shift the lens. We recognize that sustainable enterprise value is rarely derived from the model in isolation; instead, it emerges from the high-stakes architectural decisions and systemic orchestration that define its environment. As we launch our inaugural edition, we dissect a critical technological nexus: the convergence of Graph Neural Networks (GNNs), Generative AI, and the industrial-grade infrastructure of Google Cloud Platform (GCP). We argue that for complex enterprise datasets, the transition from flat vector embeddings in latent space toward non-Euclidean, graph-based relational intelligence is the primary differentiator for the next generation of resilient AI applications. 1. The Scientific Foundation: Exploiting Relational Inductive Bias Traditional Deep Learning architectures, such as Convolutional Neural Networks (CNNs) for images or Transformers for text, primarily operate on data structured as sequences (Euclidean space). While exceptionally powerful, these structures often fail to capture the topological nuances of real-world systems like supply chains, molecular structures, or fraudulent transaction webs where data is inherently non-Euclidean. Graph Neural Networks (GNNs) provide a framework for learning from data represented as nodes and edges. Unlike standard neural networks that process inputs in isolation, GNNs utilize a Message Passing paradigm. In this process, a node's internal representation is iteratively updated by aggregating information from its immediate neighbors. Instead of looking at a data point as a single row in a database, the GNN looks at who that data point "talks to" and how those connections define its identity. By utilizing Graph Attention mechanisms, we can fu

When 150,000 tickets went on sale for The Odyssey in 70mm IMAX, they sold out almost instantly. But plans change, cancellations happen, and good seats randomly open up at odd hours. To solve this, Andrew Baker from Temporal built IMAXXING : a service that monitors every 70mm IMAX showing across the US and alerts subscribers the moment great seats become available. What started as a fun weekend project quickly scaled, now over 9,000 users. I sat down with Andrew to break down the architecture: how durable execution keeps long-running workflows alive, how to debounce alerts so you don't spam users, and how serverless workers on Google Cloud Run handle sudden spikes in demand without overprovisioning. What's in the video Durable execution 101: How Temporal allows you to rewind history to the point of failure. The Entity Workflow pattern: Why there is one persistent workflow per user subscription and separate monitoring workflows per showing across the country. Signals & smart debouncing: How showing workflows send signals to wake up subscription workflows, and how a 60-second in-workflow timer batches multiple theater alerts into a single digest—without consuming active CPU while sleeping. Serverless workers on Cloud Run : How running Temporal workers as serverless containers lets compute autoscale directly with task queue depth rather than generic CPU metrics. AI agents for ops: How modern coding agents paired with Terraform and the gcloud CLI accelerated the deployment and operational dashboard setup. The point that stuck with me is how durable execution fundamentally changes how you think about long-lived state and retries. Instead of building complex cron jobs, custom retry databases, and alert queues, the workflow state itself is the queue and the timer. Have you experimented with entity workflows or running workflow workers on serverless infrastructure? How do you handle debouncing and noisy downstream APIs in your own apps?

If you have ever wanted to ship an AI-powered application without managing GPUs, model servers, or scaling infrastructure yourself, this guide is for you. Managed inference simply means letting a cloud provider run the AI model for you: you send a request, the platform handles the compute, and you get a response back. On Google Cloud, the cleanest way to do this today is to pair the Gemini Enterprise Agent Platform (formerly Vertex AI) with Google Cloud Run , dividing responsibilities between the two services. The Agent Platform serves as the orchestration and intelligence engine, while Cloud Run hosts your custom application logic, front-end UIs, or Model Context Protocol (MCP) servers. By the end of this article, you will be able to: Explain the hybrid architecture and why each layer exists Define an AI agent in code using the Agent Development Kit (ADK) Deploy your app layer to Cloud Run with a single command Choose between online and batch inference for your workload Secure and monitor the whole setup in production New to the underlying concept? Start with Google Cloud's primer: What is AI inference? Prerequisites To follow along hands-on, you will need: A Google Cloud project with billing enabled The gcloud CLI installed and authenticated Python 3.10+ and the ADK installed ( pip install google-adk ) You can also read this purely as an architecture walkthrough; every step is explained, not just shown. 1. The Architectural Blueprint This pattern splits your system into independent, auto-scaling tiers: [ Client / Web UI ] ──> [ Cloud Run Service ] (App Logic / Tool Front End) │ ▼ [ Gemini Enterprise Agent Platform — Agent Runtime ] (Orchestration, Intent Analysis, Memory) │ ▼ [ Managed Inference / Model Garden ] (Gemini 3.x Pro / Flash models) Why split it this way? Each tier scales independently and fails independently. Your web front end can handle a traffic spike without touching the model layer, and you can swap models without redeploying your application code

Nota: ✋ This post was originally published on my blog wiki-cloud.co Introduction Artificial intelligence is evolving at an unprecedented pace and is transforming how people and businesses interact with technology. Over the past few years, much of the focus has been on generative AI models, which can create text, images, code, audio, and other types of content from natural language instructions. These capabilities have marked a significant and transformative shift in how we perform many tasks, allowing AI to move from a specialized technology to an accessible tool for millions of users. However, we are entering a new stage. Artificial intelligence models are no longer limited to simply answering questions or generating content. They can now be autonomous, understand objectives, analyze context, decide what steps to take, use tools, consult different sources of information, connect with APIs, execute actions, and collaborate with other specialized agents to complete more complex tasks. This evolution is giving rise to what is known as agentic artificial intelligence, an approach in which AI systems can act with a greater level of autonomy and actively participate in business, technical, and operational processes. Instead of simply offering a recommendation, an agent can search for information, validate data, coordinate different activities, and execute a sequence of actions aimed at achieving a specific goal. Within this new scenario appears Google Agent Development Kit , also known as Google ADK , is an open-source framework developed and designed by Google to facilitate the creation, evaluation, and deployment of artificial intelligence agents. ADK provides developers with a structure for defining agent behavior, connecting them to language models and external tools, managing sessions and memory, coordinating multi-agent systems, and evaluating their performance before deploying them to production. Thanks to this code-based approach, Google ADK allows you to build e

Google just made multi-region Cloud Run significantly easier. Here is the full picture; what changed, what it means in practice, and how to build it right. Most teams discover they need multi-region architecture the hard way and sadly, during an outage. Whether you're running a global e-commerce platform, a real-time gaming API, or a financial services application, users expect your service to be available whenever they need it. There is a conversation that happens in almost every engineering team at some point. It usually starts with a post-mortem. A regional Google Cloud outage or a Cloud Run service that hit a cold start spike, or a single-region deployment that could not handle the latency demands of users spread across Lagos, Nairobi, and London simultaneously, caused enough pain that someone finally asked: why are we only deployed in one region? The answer is usually one of three things: it felt complex, it felt expensive, or no one had prioritised it yet. In July 2026, Google moved Cloud Run Service Health to General Availability and the timing was hard to miss. Six days earlier, a power cut at Google's Netherlands data centre had knocked three services offline. The GA release brings automatic cross-region failover to Cloud Run with what Google describes as a two-step setup: add a readiness probe, set minimum instances to at least 1. The load balancer does the rest. This article covers the full architecture, what Service Health is, how readiness probes underpin it, how to set up the Global Load Balancer correctly, and how to test that failover actually works. It also covers the production details. What changed: Service Health and readiness probes Before Service Health, multi-region Cloud Run required you to implement a /health endpoint in your application and configure a separate HTTPS health check at the load balancer level. This worked, but it had a significant gap. The load balancer's health check only knew whether the Cloud Run service endpoint was respon

Measured 2026-07-21 on vllm/vllm-tpu:nightly (vLLM 0.23.1rc1.dev1076), a GCE flex-start ct6e-standard-1t (one TPU v6e chip, 32 GB HBM) in europe-west4-a. TL;DR The plain google/gemma-4-E2B-it serves well on one v6e chip; none of its QAT siblings load at all. The 2-billion-parameter "efficient" Gemma 4 sustains 213 tok/s for a single user with a 16 ms first token, scales to ~2,200 output tok/s across concurrent streams, handles OpenAI-style function calling — including parallel calls and refusal to hallucinate calls — without a miss, and answers simple vision questions accurately in ~200 ms. The QAT variants are a different story: the int4 compressed-tensors export hits an unimplemented quantization path, and the bf16 QAT export trips a loader bug — the Gemma 4 implementation demands per-layer norms that E2B's KV-sharing architecture legitimately doesn't have. Filed upstream as tpu-inference #3225 . One capability coupling to know about: with a reasoning parser configured, schema enforcement only engages when thinking is enabled — thinking-off requests sail through unconstrained with a 200 status. Config interaction, not TPU limitation; details below. 1. Getting to a serving endpoint The host is a GCE flex-start VM — capacity granted on request, billed until deleted, hard-stopped at a 4-hour max run, $1.35/chip-hour. A startup script installs Docker, pulls vllm/vllm-tpu:nightly , fetches the Hugging Face token from Secret Manager via the metadata server, and launches vLLM. Boot timeline: VM RUNNING at t+0 (200 GB boot disk — the 10 GB default cannot hold the vLLM image) → Docker installed ~t+1:00 → image pulled ~t+6:00 → weights downloaded, XLA compiled, health green ~t+8:30. Two environment quirks worth knowing: Direct SSH silently times out on some networks even when the VPC allows tcp:22 — the block is upstream of the VPC. IAP tunneling ( gcloud compute ssh --tunnel-through-iap ) rides over HTTPS and works; so does tunneling the API port with gcloud compute start-

Introduction At recent technology conferences, one topic has caught my attention: every year, more health-focused devices, sensors, and applications appear. Smartwatches track heart rate, smart scales measure body data, glucose monitors record blood sugar levels, and apps help users track sleep or nutrition. Today, the amount of information we can collect about our own bodies is enormous. This article was inspired by an everyday experience with my father, Herminio ❤️ . Whenever he has a medical appointment, he opens the Apple Health app and shows the doctor the evolution of his heart rate, physical activity, sleep hours, and other recorded metrics. While watching this, I kept asking myself the same question: are we really making the most of all this information? Showing a chart during a medical appointment can be useful, but the data could provide much more value if it were automatically processed, summarized, and transformed into a structured health report. For this reason in this project, I use Gemini to transform previously calculated metrics into a clear and organized summary. The LLM does not analyze all the raw records or perform the main calculations. The pipeline processes the data, calculates the indicators, and generates the visualizations, while the model acts as a support layer for building the report narrative. The goal is not to create a medical application or replace professional judgment. Instead, the purpose is to build a prototype that shows how Apple Health exports, deterministic data processing, visualizations, and an LLM can be combined to generate automated reports. This project was developed using simulated data from three patients, so the complete pipeline can be reproduced without using real clinical information. ✨ Why Gemini? This project uses an LLM to transform previously processed metrics into a structured narrative that can be reviewed more easily by a healthcare professional. I chose Gemini for practical reasons: 〰️ I was already famil

El 12 de junio de 2026, Google Cloud publicó el Open Knowledge Format (OKF) , una especificación abierta que intenta resolver un problema que suena aburrido pero es carísimo: cómo darle a un agente de IA el contexto que necesita para no inventar. La propuesta es tan simple que da un poco de desconfianza —una carpeta de archivos Markdown con un encabezado YAML— y esa simpleza es, al mismo tiempo, su mayor virtud y el blanco de todas las críticas. Vale la pena entender qué anuncian, porque detrás del formato aparentemente trivial hay una apuesta bastante ambiciosa sobre cómo van a compartir conocimiento las empresas en la era de los agentes. El problema: el conocimiento vive en silos En casi cualquier organización, lo que un modelo necesita saber está desparramado y encerrado en formatos incompatibles: catálogos de metadatos con APIs propietarias, wikis internas, comentarios de código, docstrings, celdas de notebooks y —el clásico— la cabeza de dos o tres ingenieros senior. Cuando un agente tiene que responder algo tan concreto como "¿cómo calculo los usuarios activos semanales a partir del stream de eventos?" , tiene que ensamblar la respuesta juntando pedacitos de superficies que no se hablan entre sí. El resultado: cada equipo que arma un agente resuelve el mismo rompecabezas desde cero, y el conocimiento queda preso del sistema que lo generó. No hay portabilidad. La propuesta: un formato, no una plataforma La respuesta de Google no es "otro servicio de conocimiento en la nube" —y ese es el punto que más recalcan—. Es un formato . OKF v0.1 representa el conocimiento como: Solo Markdown : legible en cualquier editor, renderizable en GitHub, indexable por cualquier buscador. Solo archivos : se transporta como un tarball, se hospeda en cualquier repo git, se monta en cualquier filesystem. Solo frontmatter YAML : campos consultables como type , title , description , resource , tags y timestamp . Cada "concepto" (una tabla, un dataset, una métrica, un runbook) es un arc

Building an explainable AI platform that helps district administrators allocate resources and farmers make better crop decisions using Gemini, Vertex AI, BigQuery, and Google Cloud. Climate disasters are not just weather events. They are decision problems. When forecasts predict a strong El Niño, governments do not simply need more data. They need answers to questions like: Which districts will be affected first? Where should limited water resources be sent? Which crops are likely to fail? What should farmers sow instead? Why is the AI recommending this action? Existing dashboards provide plenty of charts. Very few provide decisions. That became the motivation behind El Niño 2026 Decision Copilot , an AI-powered decision intelligence platform built during the Google Cloud Gen AI Academy APAC Hackathon . The Problem India depends heavily on the monsoon. A severe El Niño can lead to: Rainfall deficits Reservoir depletion Groundwater stress Crop failures Rising food prices Rural employment challenges The information already exists across dozens of government portals, weather services, satellite datasets, and agricultural reports. The challenge is that it is scattered. District collectors do not have time to manually combine: Weather forecasts NDVI satellite imagery Reservoir levels Mandi prices Contingency plans Drought indicators Farmers face an even bigger challenge. Most need a simple answer: Given my district, should I plant the usual crop this season? The Goal Instead of building another dashboard, I wanted to build an AI system that reasons over multiple data sources and produces explainable recommendations. The platform serves two audiences through the same intelligence engine. District Administrators They receive: District risk scores Interactive risk maps Reservoir outlook Crop stress indicators Resource allocation recommendations AI-generated explanations Instead of simply showing that a district has high risk, the system explains why . Farmers Farmers intera

Even the most advanced enterprise systems are tethered to a costly paradox: manual bottlenecks that introduce critical errors, security risks, and slow innovation. These hidden operational anchors are the friction preventing your organization from realizing its full potential. The Challenge: Manual Bottlenecks in Modern Enterprise Operations In an era defined by cloud-native architectures, microservices, and declarative infrastructure, a persistent and costly paradox remains at the heart of enterprise operations. We have built systems capable of immense scale and resilience, yet they are often tethered to manual, human-driven processes that act as operational anchors. These bottlenecks aren't just minor inefficiencies; they are critical points of failure, introducing latency, human error, and security vulnerabilities into our most important workflows. They represent the friction that slows down innovation, drains resources, and prevents organizations from realizing the full potential of their digital investments. Before we can orchestrate an autonomous workspace, we must first dissect the anatomy of these manual constraints. Identifying the High Cost of Manual Invoice Reconciliation To ground this challenge in reality, consider a ubiquitous and deceptively complex business process: accounts payable invoice reconciliation. On the surface, it seems simple. In practice, it's a classic example of a high-friction, manual workflow that silently bleeds enterprise resources. The typical process is a gauntlet of context-switching and swivel-chair integration: An invoice arrives, often as a PDF attached to an email, with no standardized format. A finance professional must manually open the document and visually identify key data points: invoice number, date, vendor, line items, and total amount. They then pivot to an ERP system like SAP or NetSuite to find the corresponding Purchase Order (PO). Next, they might need to access a separate logistics or warehouse management syste

Cost is everything. In just about every agentic conversation, the three things that come up for enterprises implementing AI workloads are: Cost Observability Security and as AI continues to throw everyone for a loop when it comes to cost management (e.g - Uber running out of the yearly token budget in one quarter), the ability to shrink resource (like hardware) usage will be crucial moving forward. In this blog post, you will learn how to cust costs by 90% using Agent Susbtrate in comparison to Agents running in k8s Deployments/Pods. The Cost Comparison Agents need a place to run. The "place to run" needs to be a platform that's easily managed, orchestrated, and has the ability to cluster resources. Resources like CPU, GPU, and memory need to be able to scale and expand. Without this, it's a matter of manually managing servers that Agents are running on and clients to interact with said server. That's why so many organizations choose Kubernetes to run Agentic. When running Agents per Pod, however, that can get costly very quick in terms of hardware (GPU, CPU, memory) and performance (can your cluster scale up and down quickly based on resource needs when it comes to Agents coming up and going down per use?). The tests in this blog post show: Always-on Agents running in k8s. Actors running in Workers via Agent Substrate And the comparison will be 50 always-on Pods in comparison to 50 Actors across 5-7 Workers (Pods). If there are 50 Agents running per Pod and 50 Agents running per Worker with 5-10 Actors per Pod, you can already imagine the hardware resource savings that can be accomplished. Right now, the majority of organizations start off with the "one Agent per Pod" approach as that's the fastest way to show value and get up and running. For the future, however, Agents in Actors via Agent Substrate will be how organizations deploy when they care about efficiency, optimization, and managing cost. Let's dive in from a hands-on perspective. Prerequisites To follow a

Anatomy of the Trust Gap Before we can talk about the solution, we need to talk honestly about how...

When managing Shared VPCs, most teams allocate dedicated IP subnets for each service project to keep firewall rules simple, but this isolation often leads to poor IP utilization — it is not uncommon to see subnet IP utilization hovering in the low teens. On the other hand, using large shared subnets requires coordinating workload deployments to ensure there is enough internal IP address space for everyone. To optimize these shared networks, you need real-time visibility. The WITH_UTILIZATION query parameter on the Method: subnetworks.list | Compute Engine API solves this by returning the exact count of allocated and free IP addresses for each subnet IP range. This capability is designed for query-time decisions. For example, if you need to deploy a GCE workload requiring 100 instances, you can search for a subnet with enough capacity. This query-time data comes directly from Google Cloud's internal IP allocator and includes both primary and secondary CIDR ranges. Automating the search with gcloud and jq To automate capacity checks before you deploy, you can script this check. The script below uses gcloud compute networks subnets list | Google Cloud SDK to grab the utilization data as JSON, and then uses jq to parse, filter, and sort the subnets based on your required capacity: #!/bin/bash # --- Configuration (Replace with your details) --- PROJECT = "<YOUR_PROJECT_ID>" NETWORK_NAME = "<YOUR_VPC_NETWORK_NAME>" REGION = "<YOUR_REGION>" REQUIRED_IP_CAPACITY = 100 echo "Searching $NETWORK_NAME in $REGION for subnets with >= $REQUIRED_IP_CAPACITY free IPs..." echo "------------------------------------------------------------------------" # Fetch subnets with utilization data, output as JSON, and pipe to jq gcloud compute networks subnets list \ --project = " $PROJECT " \ --network = " $NETWORK_NAME " \ --regions = " $REGION " \ --view = WITH_UTILIZATION \ --format = json | \ jq -r --argjson min_ips " $REQUIRED_IP_CAPACITY " ' [ .[] | { name: .name, cidr: .ipCidrRange, #