Ivo Tonioni — AI Systems Engineer
I build AI systems that connect models, data, tools, and real operational constraints — from agentic workflows to leakage-aware forecasting pipelines.
Capability signal
01 / Agentic systems
AI becomes useful when it can retrieve trusted context, use the right tool and show how it reached an answer.
Architecture → action → evaluationAgent workflows
WhatsApp automation and chatbot prototypes using Meta/Baileys, Llama 3.2 and OpenAI APIs — designed around flows, triggers, external context and reliable hand-offs.
Retrieval + grounding
FAQ assistant work integrating an OpenAI agent with vector retrieval, plus local-LLM experiments for interactive data products.
AI reliability
Data and concept drift monitoring prototypes with Evidently AI and Alibi Detect, paired with synthetic-data and visual analysis tools.
02 / Forecasting lab
A complete DK2 electricity-price pipeline that moves from market data to a measurable decision layer.
forecast horizon
engineered features
research notebooks
test modules
03 / Evidence layer
A profile shaped by model development, production-minded delivery and critical infrastructure—not coursework alone.
Education
UNIR, 2024. Officially assessed in Denmark at civil-engineering master’s level (one year).
Applied ML
Streamlit data products, drift monitoring, synthetic data, local LLMs and OpenAI API prototypes.
Domain edge
Hands-on HV/MV and protection & control work on Energinet projects in Denmark.
Delivery
Foreman ownership, technical coordination, documentation and on-time field delivery.
04 / Working profile
I combine an MSc in Artificial Intelligence, electromechanical engineering, ML product work and leadership on critical Danish energy infrastructure.