Building the systems behind a multi-tenant small-business platform.
Jakomu combines public web experiences with a guided Showroom, AI-assisted client delivery, CRM-ready lead research, and shared platform infrastructure. My work spans the product surfaces, tenant boundaries, data quality, and cost decisions behind them.
Platform, Showroom, AI delivery, and lead-data engineering
State
Public surfaces plus private internal systems
Boundary
Source and operating data remain private
Live product window
Inspect a real public surface without leaving the case study.
This guarded preview loads the current Jakomu site in place. It is company-managed professional work; the surrounding case study identifies the systems and product areas I built.
For the full responsive experience, open the live site in a new tab.
www.jakomu.net / live public surfaceCompany work
↗Live preview available
The public page stays visually contained inside renaldomusto.com.
Poster-first loadingVisual preview / direct link retained
02 · The problem
What had to become clearer.
The company needed more than a brochure site. It needed repeatable client delivery, a fast path from customer inputs to tailored websites, normalized lead data for its CRM, and an architecture that could support multiple clients without duplicating infrastructure.
Constraints in the room
Company code lives in company-managed GitLab repositories and cannot be republished through a personal account.
Outbound sales remains gated by telemarketing licensing, so software readiness and operating readiness are described separately.
Lead-generation claims need to stay grounded: no inflated ROI numbers, no pretending mock/demo data is production volume.
03 · My role
The boundary of my contribution.
Software Engineer responsible for core Showroom brand controls, the tenant-isolated platform foundation, AI-assisted website-refresh workflows, and substantial portions of the AWS-first lead data pipeline. Source remains private because the work lives in company-managed GitLab repositories and cannot be published.
01
Implemented core Showroom controls for palette, theme, typography, scale, and live brand preview state.
02
Implemented the shared NestJS/PostgreSQL foundation with a central business registry and isolated tenant databases to simplify deployment and maintenance.
03
Built an AI-assisted website-refresh workflow that combines generated prototypes, reusable interface assets, visual review, and human QA.
04
Built the AWS-first lead data path and normalized 5,273 business records for HubSpot with company/domain deduplication and website quality enrichment.
05
Benchmarked Google and AWS place-data approaches using observed runs and public API pricing, then documented the selected provider boundary and data-quality tradeoffs.
06
Documented provider, data-quality, and compliance boundaries so research records remain distinct from outreach readiness.
04 · Key decisions
The choices that shaped the product.
Each decision connects an interface behavior to a system or operating constraint.
01
Make brand decisions tangible
The Showroom turns abstract brand conversations into concrete, comparable interface directions.
02
Make client delivery repeatable
A shared multi-tenant foundation and AI-assisted refresh workflow turn known customer inputs into tailored sites with visual checks and human QA.
03
Normalize research for the CRM
AWS place records are deduplicated, enriched, and normalized into a HubSpot-ready schema while manual and compliance checks remain explicit.
04
Keep cost proportional to stage
A shared backend, central registry, and per-business operational databases preserve tenant boundaries without multiplying infrastructure prematurely.
05 · Product experience
The interface is part of the evidence.
Real product surfaces and bounded interaction studies make the workflow visible without exposing private source or pretending roadmap work is live.
01
Preview
Turn brand direction into a shared visual reference.
02
Deliver
Turn customer inputs into tailored sites with visual and human QA.
03
Research
Normalize and enrich stored business data for HubSpot.
04
Verify
Separate working software from operating readiness.
Surface 01
Sanitized internal lead-research workflow, captured in local E2E mock mode.
06 · System
Cost-aware multi-tenant delivery foundation
01Public + Internal Surfaces
Public and internal surfaces
Jakomu site
Showroom
AI-assisted refresh workflow
Lead-generation workspace
02Shared Application Layer
Shared application layer
Next.js frontends
Shared Nest backend
Owner/admin routes
Public storefront routes
03Tenant + Operations Data
Registry and tenant data
Business registry
Tenant DB: jakomu
Tenant DB: rosarios
Tenant DB: client N
04Launch Controls
Launch and compliance controls
Research-only mode
DNC/license gates
Manual confirmation
Cloud spend deferral
The architecture preserves tenant boundaries while keeping deployment and maintenance centralized.
The lead-generation flow separates provider data, normalized CRM records, website evidence, and outreach-readiness checks.
07 · Verification and boundaries
Trust comes from what can be checked.
Private company code, customer data, and operating material remain private. Public evidence is limited to live surfaces, sanitized captures, system diagrams, and documented test and build coverage.
Verification path
Browser screenshots of public and internal surfaces.
Existing repo test and build scripts across app packages.
Launch-readiness docs covering DNC/license constraints and provider boundaries.
Runtime model
Jakomu repo docs describe a shared NestJS backend, central business registry, and per-business PostgreSQL databases for tenant isolation.
The pipeline uses AWS Places Stored as the durable source, normalizes records for HubSpot, deduplicates companies and domains, enriches website quality, and retains manual verification and compliance gates.
The monorepo exposes unit, coverage, accessibility, and Playwright browser gates across admin, Showroom, the Jakomu site, lead generation, and client-delivery surfaces.
The work established live public products, paying client delivery, a reusable tenant-aware foundation, and a CRM-ready data pipeline. The case study keeps company source and customer data private while making my engineering scope explicit.
Continue the story
Move With Musto
Client delivery under data, compliance, and operational constraints.
