SOFTWARE ENGINEER II - AI

Building the systems behind intelligent products.

Pedro Rocha is a Software Engineer in Talkdesk’s AI Engineering team, designing high-scale, cloud-native services for enterprise AI applications. He combines robust backend engineering with applied AI research in multi-agent systems and retrieval-augmented generation.

View GitHub ↗

Talkdesk AI Engineering · MSc AI Engineering at ISEP · Co-founder, Scooli

01 / PROFILE

Pedro works across the infrastructure, architecture, and intelligent workflows that allow AI applications to perform reliably at scale. His work spans cloud-native microservices, asynchronous systems, observability, retrieval-augmented generation, and agentic workflows.

  • Current role

    Software Engineer II - AI, Talkdesk

  • Research

    AI Agent Interoperability

  • Focus

    Reliable AI systems at scale

02 / EXPERIENCE

Current

Software Engineer II - AI

Talkdesk

May 2025 - Present · Porto, Portugal · Remote

Part of Talkdesk’s AI Engineering team, building high-performance microservices and backend infrastructure for AI and data-heavy enterprise applications.

  • Quarkus
  • Kubernetes
  • Kafka
  • RabbitMQ
  • OpenTelemetry
  • CI/CD
  • Builds and optimises scalable Quarkus microservices deployed on Kubernetes.
  • Works with Kafka and RabbitMQ to support decoupled, high-throughput asynchronous communication.
  • Uses OpenTelemetry for distributed tracing, logging, metrics, and service-health visibility.
  • Contributes to architecture and system-design decisions with an emphasis on modularity and long-term maintainability.
  • Strengthens production reliability through testing strategy and CI/CD automation.

  • Back End Developer

    Porto Editora

    October 2023 - May 2025 · Porto, Portugal

    Developed critical information systems and backend services for Portugal’s largest publishing group, including logistics, document archiving, and automated email workflows serving thousands of users.

    • Java EE
    • Oracle
    • SQL
    • Web Services

  • Junior Full-stack Developer

    Roboyo

    September 2022 - May 2023 · Porto, Portugal

    Contributed to automation and robotics management applications in an Agile environment, improving system functionality and user experience.

  • Junior Full-stack Developer

    WeMake

    May 2022 - September 2022 · Porto, Portugal

    Supported the development of responsive IMS features and their integration with backend logic.

03 / SELECTED WORK

Co-founder · AI platform in active development · August 2025 - Present

Scooli is an AI platform for Portuguese teachers that helps automate lesson plans and tests while keeping content aligned with national curriculum standards.

Pedro leads development of the platform’s specialised RAG engine: a Quarkus and Docker-based system designed to generate curriculum-aligned educational resources with strong grounding in official documentation.

  • RAG
  • Quarkus
  • Docker
  • Vector Databases
  • Microservices
  • AI Prompting
Visit Scooli ↗View GitHub ↗

System flow

  1. 1
    Official curriculum documentation (AE)
  2. 2
    Semantic retrieval
  3. 3
    Grounded AI generation
  4. 4
    Editable teacher resource
quarkusvector storedockergrounded generation

  • Engineering note

    Grounded retrieval

    The RAG pipeline prioritises official Portuguese curriculum documentation to reduce hallucinations and keep responses aligned with national educational standards.

  • Engineering note

    Semantic search

    Core infrastructure uses Quarkus and vector databases to search across complex pedagogical datasets.

  • Engineering note

    Deployable architecture

    Docker containerises the microservices architecture for consistent development environments and future scaling.

  • Engineering note

    Reliable persistence

    The data layer is being designed to manage structured user data alongside unstructured curriculum content with integrity and reliability.

04 / RESEARCH

Pedro is completing an MSc in AI Engineering at the Instituto Superior de Engenharia do Porto (ISEP), with research focused on multi-agent systems, AI agent interoperability, and standardised communication protocols.

The research investigates how intelligent agents can communicate and coordinate more effectively across systems — an essential challenge for building capable, composable AI architectures.

Programme
MSc AI Engineering
Institution
Instituto Superior de Engenharia do Porto (ISEP)
Period
July 2024 - July 2026
Research focus
AI Agent Interoperability
Conceptual illustration of interoperable agent systems.
05 / CAPABILITIES

AI systems

  • Artificial Intelligence
  • Agentic Workflows
  • RAG
  • LangChain
  • Vector Databases
  • AI Prompting

Architecture and backend

  • Microservices
  • Back-end Development
  • Quarkus
  • FastAPI
  • Jakarta EE
  • J2EE Web Services
  • REST APIs
  • Web Services

Distributed systems and operations

  • Kubernetes
  • Docker
  • Kafka
  • RabbitMQ
  • CI/CD Pipelines
  • Grafana

Data and platform

  • Oracle
  • SQL
  • Laravel
06 / EDUCATION

  • Master of Science — AI Engineering

    Instituto Superior de Engenharia do Porto (ISEP)

    July 2024 - July 2026

  • Bachelor of Science — Computer Engineering

    ISPGAYA

    October 2020 - July 2023

07 / ABOUT
Portrait of Pedro Rocha outdoors.
● online

Pedro is a Portuguese software engineer working at the intersection of backend systems and artificial intelligence. His approach is grounded in dependable architecture: making complex systems observable, maintainable, and useful in the real world.

Alongside enterprise AI engineering at Talkdesk, he is completing his Master’s research in AI Agent Interoperability and co-building Scooli, an AI platform for Portuguese education.

Currently

  • Engineering enterprise AI infrastructure at Talkdesk
  • Completing an MSc in AI Engineering at ISEP
  • Building Scooli’s curriculum-grounded RAG engine

Languages

  • Português — Native or bilingual proficiency
  • English — Native or bilingual proficiency
08 / GITHUB

Explore Pedro’s public work, experiments, and technical practice on GitHub.

Open GitHub profile ↗
09 / CONTACT

Pedro is open to conversations around AI engineering, intelligent systems, backend architecture, and collaborative opportunities.

Email PedroConnect on LinkedIn

Porto Metropolitan Area, Portugal