About
Lawyer (Ecuador) and software engineer based in Tilburg, the Netherlands, currently completing an LLM in Law and Technology (Privacy & Security track) at Tilburg University. I design, build and operate AI systems for legal practice, together with the security and infrastructure they depend on, to the standards that lawyers, clients and regulators expect.
Markatzy is a trademark protection platform I designed, built and launched independently, covering the similarity engine (phonetic, typographic, semantic and visual), payments and data protection compliance. At Grace Nelson Attorneys at Law, an immigration and real estate firm in Ecuador, I am responsible for all technology: an AI paralegal that drafts a reply to every client email for attorney review, the firm's websites, and its information security and IT operations. signal-watch is a research pipeline that tracks the AI infrastructure build-out and tests an investment thesis against new data each week. Previously, at Kriptos, I supported SOC 2 Type II certification and delivered more than 37,000 lines of production frontend code.
Education and credentials: Abogado (law degree), Universidad San Francisco de Quito; Postgraduate Diploma in Cybersecurity, University of Texas at Austin, McCombs School of Business; Software Engineering Diploma and Cybersecurity Certificate, BrainStation (Konrad Group); CompTIA Security+; AWS Certified Cloud Practitioner; Financial Accounting, Harvard Business School Online.
Outside work: long-distance running (Amsterdam Marathon 2025) and mountaineering (Cotopaxi and Iliniza Norte in Ecuador, Acatenango in Guatemala).
View CV Schedule a callProjects
Markatzy Trademark Protection SaaS
Trademark protection platform, designed, built and launched independently for the Ecuadorian market. Brands are compared on four dimensions: phonetic (IPA transcriptions in Spanish and English), typographic, semantic (vector embeddings) and visual (logo embeddings). An optical character recognition (OCR) pipeline converts the IP office's gazette PDFs into a searchable registry, and Claude on AWS Bedrock drafts opposition letters with citations verified against a legal knowledge base. Organization-level data isolation, rate limiting, a versioned privacy policy under Ecuador's data protection law (LOPDP), and subscription and credit payments. More than 3,000 commits in twelve months, 800+ automated tests and SonarQube in continuous integration. Next.js, tRPC, PostgreSQL with pgvector, pg-boss.
Grace Nelson Law Legal AI & IT Operations
Responsible for all technology at a bilingual immigration and real estate law firm in Ecuador. The firm's AI paralegal reads each client email, updates the client file, consults the firm's cited knowledge base and drafts a reply for attorney approval: a median of 2.6 minutes from email to draft across 1,379 runs, for about 220 clients and 920 matters. The agent has no permission to send email, and each run can access only one client's folder. Also rebuilt the firm's website and operate its information security: least-privilege Microsoft 365 access, sandboxed agents, monitoring and off-site backups. Python, Claude, Microsoft Graph, AWS, Cloudflare.
signal-watch AI Infrastructure Research
Research pipeline for analysing the AI infrastructure build-out, from semiconductors and memory to optical networking, power and the electricity grid. Twenty collectors load SEC filings, memory prices, Taiwanese and Korean export data, grid interconnection queues and earnings-call keywords into a time-series database. Each week a written investment thesis is scored against the data, reviewed by a sandboxed Claude agent and delivered as an email report; a backtest evaluates each trading rule against a control portfolio. Python, SQLite, Datasette, systemd on AWS. For educational purposes; not financial advice.
Update Manager Enterprise Tool
Windows agent deployment orchestrator built for Kriptos Inc. It enables IT departments to deploy and manage security software across thousands of endpoints at banks and multinationals. Frontend development: 219 commits and more than 37,000 lines of production code.
La Pampa Client Commerce Platform
Online store for an Ecuadorian frozen fruit pulp producer, built and operated end to end: Next.js storefront, Medusa.js headless commerce backend, card and mobile payments, S3 file storage and an administration dashboard, deployed on Railway with PostgreSQL and Redis.
Contact
Open to roles in legal technology, privacy and security, and AI engineering in the Netherlands, Germany and Spain.
Get in touch