Full Stack Engineer · Industrial Engineer
I build software for the industries I know from the factory floor.

• 7+ years writing code, the last 3 full-time on a modern stack: TypeScript, React, Node.js/NestJS, Firebase and PostgreSQL.
• Industrial Engineer who worked inside manufacturing operations — costing, production planning, inventory and traceability — before building software full-time.
• Designed and shipped a complete B2B SaaS ecosystem from scratch, now used by real manufacturing and retail businesses.
• 3 years of fully remote, async-first work, with overlap across European, US and Latin American time zones.
• Spanish (native) · English (professional, B2).

Small and mid-sized garment manufacturers and stores often run on spreadsheets and generic tools. They don't know the real cost of a production lot until it is too late, can't trace a garment back to its fabric and supplier, and lose hours every day answering customers by chat and typing invoices by hand.

One ecosystem that follows the business from the production floor to the point of sale and the back office:

The ERP runs in three manufacturing companies (women's clothing, jeans and kidswear) and the POS has 8+ active users. Client businesses report lower production costs and higher customer retention after adopting the ecosystem — see the numbers below.

The PRO products are commercial and proprietary, so their source code is private. Case studies and demo videos are on my portfolio.

flowchart LR
  subgraph MFG [Manufacturing]
    PI["PRO INDUSTRIAL<br/>offline-first desktop ERP"]
  end
  subgraph RTL [Retail]
    POS["PRO POS<br/>web app / PWA"]
    SA["AI Sales Agent<br/>Telegram"]
  end
  subgraph BO [Back office]
    ACC["AI Accounting Agent<br/>in testing"]
    LIC["Billing & Licensing Portal"]
  end

  PI <-->|real-time sync| CLOUD[(Cloud database)]
  POS --> DB[(Per-client relational database)]
  SA --> DB
  LIC -.->|licenses · renewals · payments| CUST([Customers])

Key decisions, in general terms:

• Offline-first where the work happens. Production floors can't stop when the internet drops, so the ERP reads from a local store and syncs with the cloud in real time once it is connected.
• A transactional core for retail. The POS relies on a relational database with ACID transactions, idempotent sales and a full inventory history, and a typed monorepo keeps the API and the web app on the same contracts.
• Isolation by design. Each client runs on its own database, so one customer's data can never reach another's.
• AI that works on real data, with a person in control. Agents use function calling against domain services: the model handles the conversation, the services validate the facts, and money decisions such as payment approvals stay with a human.
• Built to be sold and maintained. Products are licensed per customer and delivered with automatic updates.

Languages & frontend

Backend & data

AI

Cloud & analytics

Architecture: offline-first · serverless · modular monolith · monorepo · software licensing · payment integrations

• Portfolio & case studies: kevin.proserv.com.co
• LinkedIn: linkedin.com/in/kevinc-proservices
• Email: kevinc@proserv.com.co

Open to work: Full Stack Engineer roles — remote-first, also open to on-site or hybrid roles in Dubai (GMT+4).