Gabriel Esteves
AI Systems Engineer
Software engineer with 5+ years building production AI and backend systems in Python. I design LLM/RAG pipelines, train forecasting and computer-vision models, and build the FastAPI services and cloud infrastructure that put them into production — turning ML prototypes into reliable, deployed systems.
Lisbon, Portugal · Open to Remote & EU Relocation
Journey
Career & Education
NOW
Exploring & Building
Self-Directed R&D
Researching LLM orchestration frameworks, agent-based systems, vector databases and semantic search, and scalable model-serving infrastructure.
2022 – Present
Software Developer
Runtime Revolution
Designing LLM/RAG pipelines and time-series forecasting models, and building the FastAPI services and AWS/Terraform infrastructure that put them into production.
2021 – 2023
Postgraduate Degree, Computer Science and Engineering
Universidade da Beira Interior
Thesis: an end-to-end computer vision object detection system for marine wildlife identification from drone imagery — built with Python, PyTorch and YOLO, benchmarking multiple architectures up to YOLOv9c at mAP50 ≈ 95.5%.
See the resulting project ↓2021 – 2022
Junior Consultant / Engineer
Capgemini Engineering
Built and maintained full-stack features with Angular and TypeScript, integrating modular UI components with backend REST APIs as part of a client-facing consulting team.
2016 – 2020
Bachelor's Degree, Computer Science and Engineering
Universidade da Beira Interior
Foundational degree in computer science and engineering.
Feb 2020 – Present
UR3 Robotic Arm — Pick & Place
ROS pick-and-place system for a Universal Robots UR3: a fixed-pose baseline and a camera-guided depth point-cloud pipeline, both running identically in simulation or on the real arm. Most recently revamped in 2026.
2021 – 2023
Marine Wildlife Object Detection
Computer vision pipeline for sea turtle identification from drone imagery — the applied research behind the postgraduate thesis above.
Model performance
0.0% mAP50
2026
AI Chatbot & Poll Intelligence Platform
LLM-driven conversational system for Portugal's 2026 presidential elections, with contextual retrieval and citation-grounded answers.
Live & grounded
Ask about candidates, polls, and platforms — every answer cites its source.
Visit botpresidencial.blogMore projects
Also shipped
Jul 2022 – Ongoing
Financial Risk Forecasting Engine
Applied ML for debt prediction and early anomaly detection — time-series forecasting with feature engineering and automated risk-scoring integration.
2026
Interview Assistant
Real-time desktop app that listens to interview audio, detects questions, and streams AI-generated answers live.
2026
Round Table — Multi-Agent AI Chat
Messenger-style app where up to 5 AI agents with distinct personalities debate ideas — each one thinks, speaks, and remembers.
Toolbox
Core Technologies
Languages
Python · SQL · TypeScript · JavaScript · Java
Data & ML
PyTorch · TensorFlow · Scikit-Learn · Pandas · NumPy · Prophet · HuggingFace Transformers · LangChain · OpenCV · FAISS · XGBoost · LLMs · RAG
Backend, Cloud & Frontend
FastAPI · REST APIs · AWS · Terraform · CI/CD · React · Next.js · Angular · Firebase
Databases
SQL Server · PostgreSQL · ChromaDB · Pinecone
LLM Backends & Robotics
Ollama · Anthropic Claude API · OpenAI API · ROS
$ whoami
gabriel@dev:~$ AI Systems Engineer — Lisbon, PT
$ stack --core
Python · TypeScript · FastAPI · React · LLMs · RAG
$ status
Open to remote & EU relocation ✓
Get in touch
Let's build something.
Open to remote roles and EU relocation, based in Lisbon, Portugal.
© 2026 Gabriel Esteves