Service
Google Cloud AI consulting
Formare AI is a Google Cloud AI consultancy in Bucharest, Romania. We design, build and deploy production AI systems on Gemini, Vertex AI, BigQuery and Cloud Run for clients worldwide, taking notebooks, prototypes and proofs of concept to deployed systems. When the right architecture is not yet clear, the 4-Week AI Strategy & Architecture Audit (from €8,000) ends in a written plan prioritised by ROI. Otherwise projects start from €8,000 or at €600–€900 per day, after a free 30-minute discovery call.
Updated
Scope
What Google Cloud AI consulting covers
We cover the whole path from architecture to a running system: choosing the right Google Cloud services for the job, building the agents, pipelines and applications, deploying them to production and handing everything over. You own the code, the infrastructure and the knowledge.
We work primarily in the Google Cloud and Gemini ecosystem, with React and TypeScript on the frontend and Python or Node.js on the backend. If Google Cloud is overkill for your use case, we say so on the discovery call.
Platform
Google Cloud services we build with
| Service | Layer | What we use it for |
|---|---|---|
| Gemini API | Models | Agent reasoning and function calling; real-time voice over WebSocket with the Gemini Live API |
| Vertex AI | Models | Embeddings, search grounding, custom models and batch prediction |
| Agent Development Kit | Agents | Multi-agent orchestration, sessions and tool calling, with A2A and MCP for interoperability |
| Vertex AI Vector Search | Retrieval | Multimodal product search at under 300 ms P99, with the Ranking API for re-ranking |
| Vertex AI Feature Store | Retrieval | Real-time feature enrichment in under 2 ms |
| Vertex AI Search | Retrieval | Retrieval-augmented generation over internal documents |
| BigQuery | Data | Natural-language-to-SQL agents, data catalogs, pipelines and agent audit logs |
| Cloud SQL | Data | Persistent agent sessions and application databases (PostgreSQL) |
| Cloud Run | Runtime | FastAPI and Express services, plus Cloud Run Jobs for scheduled work |
| GKE Autopilot | Runtime | Self-hosted platforms, such as a BigQuery data catalog |
| IAM and Secret Manager | Security | Least-privilege access, Workload Identity and secrets kept out of code |
| Cloud Trace | Observability | Distributed tracing across agent and tool calls |
Production
Taking a prototype to production on Google Cloud
Most AI projects stall as notebooks or demo apps. Getting one into production changes the work: the system needs persistent state, access control, logging, a deployment pipeline and a cost profile you can predict.
What changes
- Persistent sessions and application data in Cloud SQL instead of in-process memory
- Least-privilege access with IAM, Workload Identity and Secret Manager instead of keys in code
- An audit log of every model and tool call in BigQuery
- Containerised deployment on Cloud Run or GKE Autopilot from the first sprint
- Cost controls such as context caching, partition filtering and materialized views
Running cost is a design input from the start. In one production retail environment, six architectural decisions cut Gemini API costs by 70%; the article under “Read next” lists each one.
Compliance
Data residency, GDPR and the EU AI Act
Systems are designed for EU-region data residency, audit trails and explainability from the first sprint, not added after launch. Our article on EU AI Act and GDPR compliance for RAG systems sets out the Google Cloud architecture and the code-level controls.
This is engineering guidance, not legal advice: your counsel should confirm how the AI Act classifies your use case.
Engagement
How an engagement works and what it costs
- 4-Week AI Strategy & Architecture Audit: from €8,000, 4 weeks, fixed scope, written deliverable.
- Day rate: €600 – €900. Flexible blocks, hourly or daily.
- Fixed project: From €8,000. Scoped after the discovery call, paid in milestones.
- Retainer: From €4,500/mo. Priority access and ongoing AI engineering support.
Every engagement starts with a free 30-minute discovery call. A written proposal with scope and milestones follows within 48 hours, delivery typically runs 2–12 weeks, and the project ends with documentation, runbooks, cloud credentials and 30 days of email support. Rates are in euros and VAT may apply.
Work
Systems built on Google Cloud
DS Data Catalog
Self-hosted BigQuery data platform
Self-hosted data catalog for GCP / BigQuery teams. FastAPI + React + PostgreSQL + GKE Autopilot. AI-powered column descriptions via Gemini, data lineage via Cloud Data Lineage API, quality scoring, OAuth2, and full table/column search. No stale Confluence docs — metadata syncs directly from BigQuery.
- Full BigQuery sync
- AI column descriptions
- Data lineage graph
Senior Merchandiser AI Agent
Retail catalog intelligence — multi-agent
Multi-agent merchandiser: Gemini planner, parallel vector search and re-ranking, Feature Store enrichment.
- +15–25% search precision
- Real-time Feature Store <2ms
AI Airport Transfer Platform
George Travel — georgetravel.co.uk
Full-stack AI booking platform for a UK private driver. Gemini agentic loop with 11 custom tools handles geocoding, fare calculation, conflict-safe booking (PostgreSQL advisory locks), Stripe payments, Google Calendar, automated emails, real-time flight tracking, and weather. Interoperable via A2A protocol and MCP.
- 11 AI tools
- 0 double-bookings since launch
- A2A + MCP interop
NL → BigQuery Assistant
Conversational data analysis — multi-agent
Ask BigQuery in plain English: an orchestrator routes to schema-aware NL→SQL and Pandas agents.
FAQ
Questions about Google Cloud AI consulting
Cost, timelines and fit, answered before the first call.
What does a Google Cloud AI consultant do?
A Google Cloud AI consultant decides which Google Cloud services fit a problem, designs the architecture, builds and deploys the system, and hands it over. At Formare AI that means agents and multimodal applications on Gemini and Vertex AI, data platforms on BigQuery, and the Cloud Run or GKE infrastructure around them, with the same team writing the code and deploying it.
Do you only work with Google Cloud?
We work primarily in the Google Cloud and Gemini ecosystem, which is where our production experience is. If Google Cloud is overkill for your use case, we say so on the discovery call.
Can you take a prototype or notebook to production?
Yes, that is the core of the work. Production usually means persistent sessions and data stores, access control, audit logging, a Cloud Run or GKE deployment and a cost profile you can predict. Cloud Run deployments happen in the first sprint, so you test against real infrastructure rather than a localhost mock.
How much does Google Cloud AI consulting cost?
Pricing is €600–€900 per day, from €8,000 for a fixed-scope project paid in milestones, or from €4,500/mo for a retainer. Rates are in euros and VAT may apply. The 4-Week AI Strategy & Architecture Audit starts at €8,000 with fixed scope and a written deliverable. The 30-minute discovery call is free, and a written proposal follows within 48 hours.
Where is Formare AI based, and who do you work with?
Formare AI is based in Bucharest, Romania, works remote-first with clients worldwide, and replies to enquiries within 24 hours on weekdays (CET). Clients range from startups and SMEs to enterprises.