Ciao, I’m
Amarildo Guga
Aerospace engineer → ML & Data Scientist
About
Data scientist and ML engineer in HSBC Payments. I started out in aerospace (BEng, First Class, Surrey; MSc Advanced Aeronautical Engineering, Imperial) and moved into machine learning.
These days I build agentic AI systems and probabilistic forecasting models, and I care as much about making them auditable and reliable as about making them clever.
Projects
Agentic test-storyboard generation
A multi-stage, multi-agent pipeline with JSON-schema contracts between stages, verification at every step and evaluation against a golden dataset.
Probabilistic cashflow forecasting
Temporal Fusion Transformers producing quantile forecasts over long horizons, with conformal calibration so the uncertainty bands can be trusted.
Payment-rejection message classification
Transformer embeddings and TF-IDF features into gradient-boosted classifiers, with HDBSCAN and UMAP to surface hidden patterns in the messages.
Server-estate automation
Ansible playbooks to install, upgrade and decommission monitoring agents across a server estate, with Vault-based SSH auth and Splunk integration.
Wind turbines in complex terrain
Imperial MSc research (Distinction): data-driven aerodynamic modelling of wind turbines sited in complex terrain.