Reliable AI
Evaluating AI agents, measuring quality, and designing feedback loops so production systems stay observable and useful.
BYOB: Bring Your Own BenchmarkLead AI Engineer - London
I am Federico Viscioletti, a statistician turned AI engineer. My work sits between modelling, product thinking, and the practical craft of turning ambiguous business questions into systems people can trust.
Evaluating AI agents, measuring quality, and designing feedback loops so production systems stay observable and useful.
BYOB: Bring Your Own BenchmarkPredictive modelling, uplift, segmentation, forecasting, and experiments shaped around decisions rather than dashboards.
Data Science InsightsBrowser-first apps, writing helpers, and prototypes that make one specific workflow lighter, clearer, or more fun.
Ink writing appI evaluate the reliability and accuracy of AI agents, with a focus on production readiness, monitoring, and trustworthy deployment.
I created and led the Data Science team, internalised a third-party model, managed two data scientists, and improved model performance with machine learning, Python and SQL automation, and custom data pipelines.
I helped develop downgrade models, built uplift modelling for discount strategies and net revenue impact, contributed to mobile strategy, and managed individuals and small teams.
I implemented Vodafone UK's first AWS-based data product, developed device launch and tariff recommendation models, built churn and upgrade propensity models, and mentored junior colleagues.
I worked in Customer Experience Analytics, building a Shiny app to measure campaign ROI from reactivation and spend lift, and reporting on early customer lifecycle behaviour.
I supported Root Cause Analysis for a banking data management project using SAS, Teradata, AngularJS, and D3, and built Bookshelf, a text-mining web app over Project Gutenberg books.
I delivered BI dashboards, SAS forecasting and financial management integrations, receipt fraud detection with image recognition and OCR, and social media data retrieval systems.
I worked as a client-facing SAS consultant, analysing data and dashboards for major Italian clients and supporting sales contracts and prospects with technical insight.
I focused on stochastic processes, machine learning, predictive modelling, data mining, decision theory, survival analysis, multivariate statistics, generalized linear models, and sampling theory.
I studied statistical analysis and inference, financial statistics, linear algebra, sampling, forecasting models, SQL, database design, Java, and programming.