Thomas Peter Griffiths

Principal Architect | AI Systems, Cloud Architecture & Technical Leadership

Email: thomas.peter.griffiths@gmail.com

Phone: +54 9 11 3018 7080

Web: tpg.com.ar

Location: City of Buenos Aires, Argentina

Profile photo of Thomas Peter Griffiths

About Me

Principal architect and technical owner of large‑scale, AI‑enabled platforms, currently leading the solution architecture of a greenfield GovTech system for the Government of the City of Buenos Aires. I design modular, deterministic workflow architectures that embed bounded AI and multimodal capabilities into auditable, production‑grade systems on AWS, and set technical direction across multiple delivery teams. As the primary interface between senior government stakeholders and engineering, I translate complex regulatory and business requirements into technical roadmaps. Backed by years of full‑stack, fintech, and enterprise‑integration experience, I communicate fluently in English and Spanish.

Job Experience

Phinx Lab — June 2025 – Present

Principal Architect (March 2026 – Present)

Client: Urban Development Agency, Government of the City of Buenos Aires (GCBA)

Stack: Node.js (TypeScript), AWS Bedrock, Bedrock AgentCore, Model Context Protocol (MCP), Docker, EC2, managed databases, Route 53

Team & Scope: Technical owner working with two additional architects and multiple delivery teams, each led by a technical lead.

  • Own the technical direction and solution architecture of a greenfield AI‑enabled platform designed to digitize and automate construction‑permit, zoning, blueprint‑review, property, and regulatory workflows.
  • Serve as the primary interface between senior government stakeholders and engineering teams, translating complex administrative and regulatory requirements into technical roadmaps, features, and delivery priorities.
  • Defined a modular workflow architecture that decomposes government procedures into reusable components, determining which stages should remain human‑operated, use deterministic software, or incorporate AI.
  • Lead the design of bounded, task‑specific AI agents embedded within controlled workflows, prioritizing determinism, auditability, maintainability, and clear module responsibilities over autonomous multi‑agent orchestration.
  • Delivered an early multimodal blueprint‑analysis module that uses vision‑enabled models to interpret technical plans and cross‑check information across blueprints and supporting documents.
  • Architect configurable code‑based workflows in which modules expose explicit preconditions and produce structured or Markdown artifacts that become part of the accumulated context of each administrative case.
  • Set architectural guardrails and make final high‑level technical decisions across teams; provide risk‑based code oversight, mentor engineers according to seniority, and refactor implementations when they drift from the target architecture.
  • Oversee AWS cloud architecture and DevOps implementation while leading MCP‑based, human‑in‑the‑loop tools that enable agency analysts to perform operational tasks through AI‑assisted interfaces.

Full‑Stack Technical Lead (June 2025 – March 2026)

Client: Ministry of Education, Government of the City of Buenos Aires (GCBA)

Stack: React.js, Node.js (TypeScript), PostgreSQL, TypeORM, Material UI

Team & Methodology: Lead a 10‑person cross‑functional team (developers, QA engineers, functional analysts) operating with SCRUM.

  • Provide technical leadership for a city‑wide academic platform used across all public schools in Buenos Aires; high‑traffic, mission‑critical environment.
  • Own delivery end‑to‑end: discovery, planning, implementation, code review, and production rollout with clear acceptance criteria and release checklists.
  • Define solution architecture, coding standards, and shared libraries; ensure consistency across React (frontend) and Node.js/TypeORM (backend) layers.
  • Collaborate with Ministry stakeholders to translate policy/operational needs into technical roadmaps, user stories, and milestones.
  • Guide database modeling and migrations in PostgreSQL; emphasize data integrity, observability, and performance under heavy workloads.
  • Drive sprint planning, backlog prioritization, and delivery predictability; facilitate ceremonies, track metrics, and remove team blockers.
  • Mentor engineers and QA; champion automated testing, code quality, and continuous improvement.
  • Emphasize reliability, maintainability, accessibility, and security suitable for a government platform.

Baufest — September 2019 – July 2025

Senior Back‑End AWS Integration Developer (August 2024 – June 2025)

Context: Nationwide retail (tire sales) integrations and payment/POS workflows

Stack: AWS Lambda (vanilla JS), Amazon S3, DynamoDB, EventBridge, CloudFormation, Datadog

Location/Mode: City of Buenos Aires, Argentina · Remote

  • Designed and developed serverless payment‑processing and POS workflows on AWS Lambda, provisioned via CloudFormation.
  • Architected event‑driven integrations leveraging S3, DynamoDB, and EventBridge to handle high‑volume retail transactions with resilience.
  • Implemented monitoring, alerting, and performance tuning in Datadog; reduced average Lambda latency and error rates on critical paths.
  • Drove adoption of best practices in code quality, CI/CD, and IaC; standardized workflows with VS Code, Jira, and Confluence to improve team velocity and collaboration.

Lead Angular Migration Developer (May 2024 – July 2025)

Client: Cencosud SA (SSE: CENCOSUD)

Focus: Legacy → Latest Angular migrations; performance and maintainability

  • Led migrations of legacy Angular apps to current versions, improving performance, maintainability, and security posture.
  • Conducted comprehensive code analysis and refactoring; decomposed monolithic modules into modern, testable units.
  • Updated and managed libraries/build tooling to ensure compatibility and reduce supply‑chain risks.
  • Collaborated with product, QA, and business stakeholders to align technical solutions with objectives and timelines.
  • Institutionalized front‑end best practices (TypeScript rigor, accessibility, error handling) to enhance UX and reliability.

JavaScript Full‑Stack Developer (July 2022 – May 2024)

Client: Telecom Argentina (NYSE: TEO) — Personal Pay (digital wallet)

Stack: React.js, React Native, Node.js, Express.js, PostgreSQL, TypeORM; Agile SCRUM

  • Developed customer‑engagement initiatives for Personal Pay, integrating front‑end and back‑end components for cohesive user experiences.
  • Built a BackOffice for loyalty program management (benefits, promotions, banners) with robust role‑based workflows.
  • Designed and implemented a rule engine to automate benefit allocation and campaign orchestration.
  • Collaborated with stakeholders to define/refine requirements; delivered incrementally with measurable outcomes.
  • Practiced CI/CD, code reviews, and testing strategies to maintain high quality across services and apps.

.NET Developer (September 2019 – July 2022)

Client: Tenaris (NYSE: TS)

Stack: .NET Framework, WCF, jQuery, SharePoint (CAML), Windows Server; Agile Kanban

  • Developed and maintained a web application for the IT help‑desk system, streamlining ticket lifecycle and resolution SLAs.
  • Implemented features to improve triage, routing, and reporting; reduced manual overhead for support teams.
  • Modernized legacy components while maintaining SLA stability for business‑critical operations.
  • Collaborated with IT stakeholders to ensure system reliability, security, and compliance.
  • Provided technical support and onboarding/training for end‑users and administrators.

Education

Bachelor's Degree in Computer Engineering

Ingenieria Informatica, Universidad del Salvador, Buenos Aires, Argentina

Certifications

Skills & Tools

  • Languages: JavaScript, TypeScript, C#, Java
  • Web & Mobile: React, React Native, Expo, Expo Router, Material UI, Astro
  • Back‑End: Node.js, Express.js, Bun, TypeORM, REST APIs
  • Cloud & DevOps: AWS (Lambda, API Gateway, S3, DynamoDB, EventBridge), CloudFormation, Datadog, Docker/Podman, Nginx, Linux (Ubuntu), GitHub Actions, EAS CLI
  • Databases: PostgreSQL, SQL Server, MySQL, SharePoint lists (CAML), basic Redis
  • Testing & QA: Jest, React Testing Library, Playwright/Cypress (familiarity), Postman
  • Agile & Collaboration: SCRUM (certified), Kanban, Jira, Confluence, Notion, Figma
  • Productivity & AI: GitHub Copilot, ChatGPT, Claude AI, Codespaces, Microsoft Office