Open to opportunities

Gonçalo Monteiro

@gonalomonteiro

Message

Machine Learning Engineer building RAG, LLM fine-tuning, and computer-vision anomaly detection systems.

Portugal

Message

What I'm looking for

I want to build AI systems that ship—RAG assistants, LLM fine-tuning, and anomaly detection—using strong evaluation and clean backend APIs. I thrive in research-driven teams where I can iterate quickly from experiments to production-ready pipelines.

I’m an AI / Machine Learning Engineer and Research Fellow at INESC TEC, focused on turning advanced models into reliable, private AI experiences. I developed a private multimodal RAG-based conversational assistant that lets users query administrator-uploaded files through a controlled knowledge base.

My work blends strong backend engineering with evaluation-driven research: I built API-based workflows using FastAPI, LangChain, ChromaDB, and Docker, and explored automated fine-tuning and graph-based RAG architectures to improve retrieval and generation quality. I also studied LLM-based recommender systems and their fit for personalised recommendation workflows.

At the University of Porto, my Master’s thesis centered on “Unsupervised Image Anomaly Detection and Localisation for Industrial Textile Inspection.” I built a benchmark pipeline to compare unsupervised model families across textile inspection datasets, evaluating localisation performance across texture patterns, inductive biases, zero-shot capabilities, and production-line suitability.

Experience

Work history, roles, and key accomplishments

Current

Research Fellow

Current

INESC Tec

Oct 2024 - Present (1 year 8 months)

Developed a private multimodal RAG-based conversational assistant that lets users query administrator-uploaded files via a controlled knowledge base. Built backend APIs and retrieval/knowledge integration pipelines, and explored fine-tuning and graph-based RAG to improve retrieval and generation quality.

fastAPI Vector Databases ChromaDB SQLite Docker Ollama Hugging Face

Curricular Research Intern

INESC Tec

Feb 2024 - Jul 2024 (5 months)

Fine-tuned and evaluated small language models for code generation under memory and hardware constraints. Worked on LLM training and evaluation workflows with a focus on performance under limited computational resources.

LLM Fine tuning Model Evaluation Memory Constrained Inference

Education

Degrees, certifications, and relevant coursework

University of Porto — Faculty of Engineering (FEUP)

Master’s Degree, Artificial Intelligence

2024 - 2026

Activities and societies: Selected coursework: advanced ML/AI, Deep Learning & RL, NLP, IR/Search, Computer Vision, Intelligent Robotics, Time Series. Thesis: unsupervised image anomaly detection & localisation for industrial textile inspection; built benchmark pipeline. Stack: PyTorch, PyTorch Lightning, Anomalib, MLflow, Git, Docker.

Master’s degree in Artificial Intelligence with coursework across advanced machine learning, deep learning, reinforcement learning, NLP, information retrieval, computer vision, robotics, and time series. Thesis focused on unsupervised image anomaly detection and localisation for industrial textile inspection, including benchmarking across textile datasets and evaluation of localisation performance

University of Porto — Faculty of Science (FCUP)

Bachelor’s Degree, Artificial Intelligence and Data Science

2021 - 2024

Activities and societies: Projects: supervised urban sound classification; AlphaZero-style RL for Go and Ataxx; multi-agent traffic simulator (SPADE); search-based game environments; Webots robotics wall-following.

Bachelor’s degree in Artificial Intelligence and Data Science, covering supervised learning model development and applied ML projects. Included work on urban sound classification, reinforcement learning (AlphaZero-style) for Go and Ataxx, multi-agent traffic simulation, search-based problem-solving environments, and robotics wall-following in Webots.