Marcos Toledo

Design Engineer | AI Design Technologist

Amsterdam, Netherlands · marcostolcom@gmail.com · linkedin.com/in/marcostolcom · marcostolcom.pro

Summary

I build the tools designers and engineers use. I came from industrial design, moved into computational design and parametric systems, and now deliver complete digital products: research, design, architecture, code and release.

I understand geometry, manufacturing constraints and design intent, and I build the software that encodes them.

Configurators, validation frameworks, automation that turns days of manual work into minutes, and consumer apps in production on both app stores.

Experience

Independent Consultant, Design Engineering

Amsterdam, Netherlands | 2026 - Present

• Built a local-first desktop application that reads an Arquímedes construction budget plus a folder of technical datasheets and returns the official sustainability certification template pre-filled, replacing days of manual cross-referencing per certified project.
• Architected a five-step pipeline (parse, match against a curated master table, AI on datasheets, AI on the long tail, write Excel), streamed over SSE with every step independently testable.
• Kept all project data on the client's own machines: a curated table of around 24 known materials resolves most line items with no model call, so the AI only sees the long tail, which caps cost, latency and wrong answers.
• Served prompts from an edge function with a bundled offline fallback, so a prompt change reaches the studio on next launch without rebuilding or reinstalling the binary.

Packhunt

Amsterdam, Netherlands | Computational Designer, 2022 - 2025 | Part-time Technical Product Owner, 2024 - 2025

• Built an automated regression testing framework for parametric Grasshopper models: JSON state snapshots, viewport screenshots for visual regressions, and nightly batch runs, so broken geometry is caught before it reaches a client.
• Led the technical overhaul of a six-typology timber structure configurator: a three-layer architecture separating inputs and validation, mathematical logic and geometry output, delivering an 80% faster viewport and BOM-ready outputs.
• Converted customer-approved documentation into machine checks, so specification disagreements surfaced as failing checks instead of meetings.
• Shipped a PDF generation plugin that composes geometry, images and parametric data into client reports through templates: report production dropped from days to hours and the whole office ships one format.
• Delivered the Grasshopper backend of a prefab roof configurator where product families are JSON data rather than code, so a new product type is added by data entry.
• Prioritised the internal tooling backlog as part-time Technical Product Owner, defining requirements through structured interviews with the computational designers who used the tools daily.
• Written letter of recommendation from the CEO, Jeroen Coenders (July 2025): "Marcos played a crucial role in the creation and maintenance of our clients' solutions."

PlaYa!

Personal product, live on the App Store and Google Play | 2025 - 2026

• Shipped a beach decision app covering 221 swimming spots across 8 Canary Islands, solo, on both stores.
• Built a real-time pipeline over three official sources (Open-Meteo, AEMET, NÁYADE) refreshed every 15 minutes.
• Designed a composite 0-100 score with weakest-link logic across time windows: one bad hour drags the score down, because a plan is only as good as its worst hour.
• Added sport modes with their own scoring models: 21 surf breaks and 11 dive sites reshape the map and the ranking.
• Owned brand, landing page and SEO. Roughly a third of installs are active users, with no marketing spend, no account and no ads.

Flarekeep

Personal product | 2026

• Shipped a local-first job search copilot covering fit scoring, tailored CV and letter, interview preparation and pipeline tracking, running on the user's machine against their own AI provider.
• Constrained generation instead of trusting it: the model fills a validated design spec, so every document compiles and a contrast guard keeps it legible.
• Set the validation gate before shipping: a public waitlist, a price, and 50 signups required. Zero signups, which answered the question at the cost of one launch instead of one year.

Education

Master's Degree in Product Design

Universidad CEU Cardenal Herrera | 2021

Engineering Degree in Industrial Design and Product Development

Universidad de Las Palmas de Gran Canaria (ULPGC) | 2018

Skills

Architecture and systems: three-layer configurator architecture, parametric data modelling, JSON snapshot testing, visual regression testing, local-first architecture, data tree management, monorepo delivery, human-in-the-loop AI, privacy-by-architecture.

Engineering: Python, C#, TypeScript, React, React Native, FastAPI, Grasshopper and Rhino (Grasshopper SDK, RhinoCommon API), Typst, LLM APIs, automated end-to-end testing.

Product: discovery, scoping, validation gates, requirement definition, stakeholder alignment, technical product ownership.

Languages

Spanish: native. English: professional working proficiency. Dutch: learning.

← Portfolio · llms.txt