Delivery proof, not a logo wall.
Each case study connects a real product surface, domain explanation, architecture evidence, and targeted articles for high-intent searches.
Public case studies
These are the projects CloudFloo can describe as owned delivery work. Sabre and Samsung stay as team experience signals, not company portfolio claims.
FIRE planning with a city and date
A financial-independence product for digital nomads and ambitious professionals. IndepAI turns "retire someday" into a concrete FI number, target city, and timeline, supported by calculators, AI coaching, geo-arbitrage data, and Supabase-backed account flows.
Access-control modernization with a safer domain core
HSK WebKD is an enterprise access-control and RCP system, not a brochure site. CloudFloo is modernizing the software around RCP servers, TKD controllers/readers, PKD cardholders, KAD access cards, individual and group permissions, schedules, events, reports, remote-work workflows, and release gates while preserving the operational behavior that facilities depend on.
Bilingual commerce for professional disinfectants
A public commerce site for Hanbud: professional disinfection products for waterworks, pools, hospitals, industry, and B2B buyers in Poland and Ukraine. The site supports product discovery, catalog download, contact flows, cart behavior, and business offer requests.
Container ordering in two minutes
An in-progress marketplace for ordering rubble containers. The product collects address, waste type, volume, phone, and timing, then routes the request to verified local operators with active BDO numbers and fast callback expectations.
Targeted articles
These articles split the case studies into search intents buyers actually use: AI product architecture, access-control modernization, B2B commerce UX, and agentic delivery.
IndepAI architecture: deterministic FIRE math with useful AI at the edge
IndepAI uses AI where explanation matters, not where exact financial calculations should stay deterministic. That product boundary is the architecture.
Read articleConsent-gated AI coach tools: how IndepAI keeps financial assistance bounded
A case-study article on AI coach boundaries: consent checks, explicit tools, PII handling, cost limits, usage logs, and deterministic financial context.
Read articleSupabase AI SaaS privacy architecture: lessons from IndepAI
How an AI SaaS product can use Supabase, RLS, deterministic data models, consent boundaries, and audit-friendly product flows before scaling AI features.
Read articleHSK WebKD: what modernization means in an access-control system
Modernizing WebKD is not about a new UI skin. It is about preserving domain behavior while improving architecture, tests, local infrastructure, and delivery confidence.
Read articleAccess-control permissions as a domain model, not a CRUD screen
A WebKD case-study article on PKD cardholders, KAD cards, TKD readers, RCP servers, groups, schedules, events, reports, and permission recalculation.
Read articleKubernetes release gates for database migrations in enterprise apps
A WebKD modernization article on Helm migration Jobs, init-container waits, probes, image publishing, Nx affected builds, and safer rollout sequencing.
Read articleHanbud B2B disinfection commerce: product discovery, catalog, and offer paths
A case-study article on a bilingual B2B commerce surface for professional disinfectants, with separate paths for product discovery, catalog download, cart, and formal offer requests.
Read articleB2B commerce trust signals for specialized products
How specialist B2B websites should use certification signals, use cases, catalog paths, delivery expectations, wholesale language, and direct contact to reduce buyer risk.
Read articlePolish and Ukrainian B2B localization is product architecture, not translation
Why PL/UA localization for B2B commerce should shape buyer paths, trust signals, product copy, catalog access, and contact flows instead of living as a language toggle only.
Read articleCloudFloo Agentic Delivery Framework: AI delivery with gates, evals, and rollback
A practical framework for using agents in delivery work without pretending the model replaces architecture, ownership, testing, approvals, or production rollback.
Read articleSafe autonomy gates for production AI agents
A practical model for deciding which agent actions can run automatically, which need human approval, and which require rollback before execution.
Read articleAutonomous AI agents need an operating system, not just a prompt
Production agents need durable memory, source-grounded retrieval, roles, approvals, audit trails, and an operational control surface before they can safely work inside company systems.
Read article