Miguel Duarte, Developer in Lisbon, Portugal
Miguel is currently unavailable

Miguel Duarte

Bio

Miguel is a hands-on full-stack and AI engineer who, as the founding CTO of Care&, built a healthcare platform from scratch into an AI-native product: a custom AI Scribe, patient-facing assistants, and agentic operations automation. On the executive team, he helped grow the business to 7-figure revenue while leading a small engineering team. He applies frontier AI across the business: engineering, QA, infrastructure costs, SEO, and ads, giving AI agents the feedback loops to iterate and ship.

Portfolio

Care & Family Health
Amazon Web Services (AWS), Express.js, Redux, Docker, React, Node.js...
Kaja (via Toptal)
Java, AngularJS, PHP, Node.js, Google Maps, MongoDB
Universidade Europeia (Laureate International Universities)
Assembly, MySQL

Experience

  • JavaScript - 16 years
  • Node.js - 11 years
  • Artificial Intelligence (AI) - 9 years
  • Amazon Web Services (AWS) - 9 years
  • React - 9 years
  • Large Language Models (LLMs) - 3 years
  • TypeScript - 3 years

Preferred Environment

Amazon Web Services (AWS), JetBrains IDE, Git, MacOS, Google Cloud, Claude Code

The most amazing...

...project I've worked on is building an AI-native healthcare platform end-to-end, from the first line of code to the AI Scribe clinicians use every day.

Work Experience

CTO

2017 - PRESENT
Care & Family Health
  • Architected and built the company's healthcare platform end-to-end as founding CTO, from the first commit to the system that runs every aspect of a multi-provider primary care practice: patient-facing apps, clinical tooling, and infrastructure.
  • Led the platform's evolution into an AI-native product: a custom AI Scribe that drafts visit notes and summarizes charts, patient-facing AI assistants for intake and booking, and an AI inbox that routes and drafts replies to patient messages.
  • Served on the executive team from inception, shaping product strategy and growth as the company scaled to seven-figure annual revenue.
  • Ran an AI-first engineering and business practice, applying frontier models to development, QA, infrastructure cost optimization, and growth (SEO and ads), and giving AI agents the tools to iterate and verify their work (CI, infrastructure, logging).
  • Built agent-based automation for back-office operations such as user provisioning, scheduling, and administrative workflows, substantially reducing manual operational work.
  • Hired, mentored, and led a small engineering team, owning technical direction, code quality, and delivery across all projects.
  • Implemented security and privacy by design for sensitive health data (PHI), with encryption at rest and in transit across all components, in line with Canadian healthcare privacy requirements (PHIPA).
  • Designed and operated the company's cloud infrastructure on AWS (ECS, S3, DynamoDB, SNS, and SQS) with Docker-based CI/CD, reliably serving 50,000+ patients while optimizing resource costs.
  • Integrated dozens of third parties, including payments (Stripe), government health systems (OHIP, Ontario Ministry of Health), HL7 clinical data exchange, eFax, telephony (Twilio, Microsoft Teams), and AI providers (Anthropic, OpenAI, Google).
  • Implemented fully-featured telemedicine capability into the system, where patients can have video appointments with a practitioner.
Technologies: Amazon Web Services (AWS), Express.js, Redux, Docker, React, Node.js, TypeScript, Artificial Intelligence (AI), Large Language Models (LLMs), AI Agents, Claude API, OpenAI API, Amazon DynamoDB, Stripe, WebRTC, Linux, Sass, JSON, REST, Swagger, Twilio API, Prompt Engineering, Socket.IO, HL7, Firebase, Redis, AWS Bedrock AgentCore, Vertex AI, WebSockets, Vite, Material UI, Amazon S3 (AWS S3), Amazon Simple Queue Service (SQS), Speech-to-Text (STT), Deepgram, Amazon Transcribe, Optical Character Recognition (OCR), Technical Leadership, Telemedicine, Electronic Health Records (EHR), HIPAA Compliance

Full-stack Developer

2016 - 2017
Kaja (via Toptal)
  • Developed an intelligent itinerary planner accounting for user preferences and point-of-interest properties (location, rating, category) using machine learning techniques such as spatial clustering and genetic algorithms.
  • Built a recommender system for events based on the users' Facebook liked pages. The system further cross-referenced the Facebook data with Last.fm's API to discern musical tastes.
  • Added i18n to the product (both static and dynamic elements) using Microsoft Azure's Cognitive Services for automatic cached content translation.
  • Implemented a customized itinerary viewer with Google Maps.
  • Developed custom map overlays and integrated with the Directions API with client-side caching.
  • Handled the complete redesign/rebranding of the website.
Technologies: Java, AngularJS, PHP, Node.js, Google Maps, MongoDB

Invited Assistant Professor

2016 - 2016
Universidade Europeia (Laureate International Universities)
  • Taught and handled the curricular unit databases given to first-year students majoring in game and app development.
  • Taught and contributed to the curricular unit computer architecture given to first-year students majoring in computer science and computer science with management.
  • Placed in the top 10% of faculty with regard to student satisfaction.
Technologies: Assembly, MySQL

Postdoctoral Researcher

2016 - 2016
Instituto de Telecomunicações
  • Developed a novel technique to automatically synthesize neural-based control for robots with complex morphologies.
  • Built a client-server application to interact with the V-REP robotics simulator via its API, enabling the parallel and distributed evaluation of robotic controllers.
  • Published three peer-reviewed publications in evolutionary robotics.
  • Won the Fraunhofer Portugal Challenge 2016 for my PhD research work.
Technologies: V-REP, Artificial Neural Networks (ANN), Evolutionary Computation, Java

Doctoral Researcher

2012 - 2016
University Institute of Lisbon (ISCTE-IUL)
  • Developed a novel artificial intelligence approach for automatic synthesis of robotic control.
  • Designed and developed the hardware, software, and artificial intelligence for a real-world swarm robotics system.
  • Published 24 peer-reviewed publications in evolutionary robotics and AI.
  • Maintained and improved a custom distributed computing system.
  • Applied digital fabrication techniques for fast, iterative hardware design.
  • Won a Paper of Excellence award (IEEE ICDL-EPIROB, 2012), was nominated for a Best Paper award (PPSN, 2016), and won a Best Robot Video award (AAAI Video Competition, 2016).
  • Wrote my PhD thesis and won the "Fraunhofer Challenge" (Fraunhofer Portugal AICOS) and the "Best PhD Thesis in Robotics" award (Portuguese Robotics Society).
Technologies: Electronics, Robotics, Machine Learning, Artificial Neural Networks (ANN), Evolutionary Computation, Java, Artificial Intelligence (AI), Robot Operating System (ROS), C, OpenCV, R, Python, Eclipse IDE, Ggplot2, Eclipse

Teaching Assistant

2013 - 2015
University Institute of Lisbon (ISCTE-IUL)
  • Taught the curricular unit: Introduction to Programming to first-year students majoring in computer science, computer science with management, and computer science with telecommunications.
  • Taught the curricular unit: Object-Oriented Programming to first-year students majoring in computer science, computer science with management, and computer science with telecommunications.
  • Organized multiple events at the university in collaboration with the local IEEE student branch.
Technologies: Object-oriented Programming (OOP), Java

Web Developer

2009 - 2012
PTWS Lda
  • Developed custom full-stack eCommerce solutions based on open-source platforms.
  • Created third-party payment gateways for multiple eCommerce platforms.
  • Managed a small team of developers and coordinated client project deliveries.
  • Installed, configured, and developed modules for web software (Joomla, PrestaShop, WHMCS, and more).
Technologies: CSS, HTML, JavaScript, PHP, PrestaShop, Windows, jQuery, RedBean ORM, ASP.NET, XML, Ajax, Adobe Photoshop, Adobe Illustrator

Experience

JBotEvolver

https://github.com/BioMachinesLab/jbotevolver
JBotEvolver is a Java-based open-source and versatile simulation platform for education and research-driven experiments in evolutionary robotics (ER). JBotEvolver has been used in previous ER studies of our research group: in offline evolution to online evolution and learning, single to multi-robot systems, and in a number of undergraduate and graduate courses at the University Institute of Lisbon (ISCTE-IUL).

Conillon

https://github.com/BioMachinesLab/conillon
Conillon is a lightweight distributed computing platform. It allows Java-based tasks to be computed on a grid of computers with minimal effort and overhead. The necessary Java classes are loaded by worker nodes as needed.

A Sea of Robots

https://www.youtube.com/watch?v=JBrkszUnms8
Won the "Best Robot Video" award at the annual video competition of the AAAI international conference. The video showcases a real-world swarm robotics system for which I designed and developed the hardware, the software, and the artificial intelligence controllers as part of my doctoral research.

Color Orbs

https://www.microsoft.com/en-us/p/color-orbs/9wzdncrdrhpf
A game for Windows Phone that tests the player's memory and reaction speed. The player has to click colored spheres in the correct order and can also use power-ups to make the task easier. Color Orbs received the first prize in the Windows Phone App Code Camp competition.

Jhip8

https://github.com/miguelduarte42/Jhip8
Jhip8 is a Chip-8 interpreter written in Java and was intended as a first experiment in the world of emulation. The project implements all the opcodes and features with the exception of sound, and it was a fun exercise in low-level programming and virtual machine internals.

Chip8JS

https://github.com/miguelduarte42/Chip8JS
This is a JavaScript port of my Jhip8 Chip-8 emulator that runs directly in the browser, using an HTML5 canvas for rendering. Porting the interpreter from Java to JavaScript was a fun exercise in translating low-level emulation logic across languages.

A working version can be found at the link below:
http://miguelduarte.pt/2012/07/09/chip-8-emulator-in-javascript/

Moppy and Floppy Guitar

I adapted the open-source project Moppy to be able to play sound in multiple floppy drives using a MIDI keyboard as input. A later version allowed music to be played on the drives using a real guitar (see URL for videos).

Education

2012 - 2016

PhD in Computer Science

University Institute of Lisbon (ISCTE-IUL) - Lisbon, Portugal

2010 - 2012

Master's Degree in Computer Science

University Institute of Lisbon (ISCTE-IUL) - Lisbon, Portugal

2007 - 2010

Bachelor's Degree in Computer Science

University Institute of Lisbon (ISCTE-IUL) - Lisbon, Portugal

Skills

Libraries/APIs

Node.js, Twilio API, Google Maps, React, jQuery, Socket.IO, OpenCV, Ggplot2, Claude API, OpenAI API, Stripe, WebRTC

Tools

PrestaShop, Eclipse IDE, RedBean ORM, Git, Adobe Photoshop, Adobe Illustrator, Claude Code, Amazon Simple Queue Service (SQS), Amazon Transcribe

Languages

Java, CSS, HTML, SQL, JavaScript, PHP, Assembly, Sass, XML, C#, C, Prolog, R, Python, TypeScript

Frameworks

ASP.NET, Redux, Swagger, AngularJS, Express.js, Material UI

Paradigms

Object-oriented Programming (OOP), REST, HIPAA Compliance

Platforms

MacOS, Linux, Amazon Web Services (AWS), Windows, Eclipse, Docker, Firebase, Vertex AI

Storage

JSON, MySQL, Redis, Amazon S3 (AWS S3), MongoDB, Google Cloud, Amazon DynamoDB

Other

Evolutionary Algorithms, Neural Networks, Robotics, Artificial Intelligence (AI), Technical Leadership, Ajax, Prompt Engineering, WebSockets, Telemedicine, Electronic Health Records (EHR), Evolutionary Computation, Artificial Neural Networks (ANN), V-REP, Machine Learning, Electronics, JetBrains IDE, Robot Operating System (ROS), Large Language Models (LLMs), AI Agents, HL7, AWS Bedrock AgentCore, Vite, Speech-to-Text (STT), Deepgram, Optical Character Recognition (OCR), Algorithms

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring