Education
I am an AI engineer with an MSc in Artificial Intelligence Engineering from ISEP, Instituto Superior de Engenharia do Porto.
Hugo Parreão
AI Forward Deployed Engineer · MSc AI Engineering, ISEP
Applied AI research, production systems, and technical writing. I work at the boundary between lab and customer: designing evaluations, hardening pipelines, shipping what comes out of the paper into the product.
About
AI engineering for systems that must operate under real constraints, explain decisions, and remain auditable.
Responsible engineering
I build AI systems with evaluation, traceability, and risk-proportionate controls, connecting engineering decisions to security and regulatory requirements.
Structured knowledge
My experience includes inference systems, ontologies, and knowledge graphs that relate entities and support automated reasoning over complex domains.
Agentic systems
I design systems with short- and long-term memory, guardrails, MCP integrations, and observability. Architectural decisions account for context, security, cost, latency, and auditability.
Python
LangGraph
FastAPI
Redis
Amazon Bedrock
Langfuse
LiteLLM
PostgreSQL
Docker
AWS
Featured projects
doclingconverter
Privacy-conscious document conversion with Docling on AWS Lambda.
Awesome-AI-Evaluation-Guide
Implementation-focused guide for evaluating LLMs, RAG systems, and agentic AI in production.
multiagentsystem-parkingmanagement
Parking management with IoT and an integrated multi-agent system.
rag-voice-assistant
RAG-powered voice assistant for document-grounded interaction.
Writing
Technical notes on applied AI: model evaluation, RAG in production, context engineering.
- The Failure of Legacy Evaluation in AI2025-11-24
- POML: A Guide to Structuring Prompts and Improving AI Interaction2025-09-02
- Context Engineering: what really improves the performance of LLMs2025-07-24
- How to Build a RAG That Processes PDFs and Answers Questions in Real Time by Voice2025-06-16
- Transform Your Productivity with 3 AI-Powered Tools2025-06-12
Threads
Short notes and takes — extended versions of my posts.
Skills
Reusable components and playbooks for taking applied AI to production.
Backend & API Hardening
Procedural memory for diagnosing and fixing recurring performance, security, and error-handling issues in backend services and APIs.
Mac Developer Cleanup
A safe workflow for identifying and removing regenerable developer artifacts on macOS, always inspecting sizes and requesting confirmation before deletion.
MCP Security Hardening
Procedural memory for designing, reviewing, and hardening MCP servers with least privilege per tool, contract validation, and safe agent-facing responses.
LGPD Engineering
A single skill for translating LGPD into engineering controls: data mapping, minimization, consent, data-subject rights, retention, vendors, and incident response.
Ultra Writing
A clear, concise, reflective writing framework for technical texts, articles, messages, and social posts, with tone calibrated to purpose and audience.
Contact
Open to applied research, mentorship, and projects. For collaborations, describe context and scope in your first message.
Talk to me on WhatsAppOr directly: hugo.aifde@gmail.com