Grounded in experience. Focused on clarity.

Virginia Levy Abulafia


AI Governance & Decision Systems | Auditability, AI Readiness, BI



For 25 years, I worked in environments where decisions are not a theoretical exercise: when you make a mistake, the best-case scenario is that everything stops.
If it goes worse, it goes much worse.

I operated in high-pressure, highly regulated environments (airport operations), dealing with complex systems, operational coordination, and distributed responsibility. That’s where I learned a simple but uncomfortable truth: complexity is not “managed”, it is engineered.

Today, I apply this approach to Business Intelligence, functional analysis, and requirements engineering, working on data-driven decision systems and on introducing AI into real-world operational contexts. My focus is not the model, but the system: data, constraints, decision flows, accountability, failure modes.

I don’t work on AI as experimentation.
I work on what happens when data and AI are put into production and start influencing operational decisions.

I intervene on:

– definition and validation of functional requirements

– data quality, consistency, and limitations

– AI readiness and data governance

– compliance-by-design processes and documentation

– decision guardrails under uncertainty

I design robust, traceable, and auditable decision systems, built to hold over time, not just to work under ideal conditions.

My position is clear, and it goes beyond AI: accuracy is a form of ethics.

Poorly structured data leads to fragile decisions.
Fragile decisions eventually turn into operational problems.
still curious?
My Tools

LANGUAGES: 

Python · R · SQL

FAVORITE LIBRARIES:

Pandas · Matplotlib · Seaborn · Folium · Scikit-learn

BI & WORKFLOW: 

Excel · Tableau · Looker Studio · BigQuery · Google Sheets

AI:

Prompt Engineering · APIs · Predictive Modelling · Compliance

Gallery
Lexie – GDPR & AI Act Compliance Copilot
AI Risk Assessment HR CV Screening System
Lexie analyzes policy documents (PDF) or free text and compares them to key articles in the GDPR and EU AI Act. It returns a downloadable human-readable PDF report with risk score, violations, citations, and recommendations.
AI Multi-Product PDF Generator – Bilingual Product Sheet Automation. It combines Generative AI (text), image integration, and layout with ReportLab.
An AI-powered compliance agent that evaluates meeting content against organizational policies.
The Hidden Cost of Unresolved Issues – Google Fiber Repeat Call Analysis
Traffic Volumes: Hours of the Day – Minnesota DOT
Agent Meme – Because crafting a clever LinkedIn comment shouldn’t take longer than reading the post
A ML project to predict whether the Falcon 9 first stage will successfully land after launch
Comparative Data Cleaning (Python & R) – Wine Quality Dataset
Designed for Desire – A Cultural and Data-Driven Journey Through Sex Toy Evolution

AI Policy

I follow a transparent and responsible approach to AI, aligned with ISO/IEC 42001 principles.
My work integrates risk assessment, human-in-the-loop design, and traceability by default.
Responsible AI is not a constraint. It is a cultural choice that enables trust and long-term value.

go to projects
Scroll to Top