I build production AI systems, end to end.
Architecture, machine learning, backend, frontend and infrastructure — I ship across all of them. An MSc in Data Science (110 e lode, University of Verona) on top of a mechanical and robotics engineering background, which is why what I build is shaped by real industrial constraints: on-premise deployment, confidential data, auditable decisions.
Delivered at Téchnéos, an AI solutions consultancy where I work as CTO and remain a primary contributor on every codebase. Client names and internal product names are omitted — what is described is the engineering.
A reusable, self-hostable foundation for enterprise AI applications: identity, fine-grained authorization, AI gateway, governed tool execution, durable workflows, audit and observability — so each client project writes domain logic only.
Digital-twin and design-of-experiments platform that lets process engineers plan experiments and predict outcomes instead of running trial-and-error batches. Includes a distillation soft sensor with calibrated prediction intervals.
Probabilistic demand forecasting with quantile regression, feeding automated safety-stock and restocking decisions under a target service level. Four model families compared per series, with forecast health monitoring.
Indexes a manufacturer's PFMEA and 8D quality documents, then searches, chats, cross-references and drafts from them — running entirely inside the customer network with local models and hybrid retrieval.
Remote GPU-accelerated OCR plus schema-validated LLM extraction, turning dense technical documents into standardized JSON/CSV datasets — with a visual schema editor and a multi-run evaluation benchmark.
Supplier invoices in, ISO 20022 payment file out: OCR, LLM extraction, master-data validation, and a human supervision desk for everything uncertain. Crash-resumable workflows, full audit trail, no automatic money movement.
On numbers: everything here is written without client names, product names or customer data. Figures are either measured and logged in the repository, or explicitly labelled as engineering estimates on each page.
Projects from my MSc in Data Science and personal research. The code for these is public on GitHub.
State-based market modelling over 20 years of equity data: transition matrices built from EMA and RSI regimes, expected-value scoring and threshold-driven trading — benchmarked honestly against buy-and-hold, which won.
Collaborative filtering and content-based recommenders over 3.6M user interactions from the Million Song Dataset — matrix factorization, implicit feedback modelling and hybrid strategies.
Transfer-learning comparison across CNN baselines, EfficientNetB0, MobileNetV2 and Xception, reaching 98.75% accuracy — including the batch-size and memory trade-offs that made Xception trainable.
I like problems where the model is only half the work, and the other half is making it dependable, deployable and explainable.
As CTO I own architecture and standards across the AI portfolio while staying hands-on — typically 50–60% of commits on the projects I lead, across backend, frontend and infrastructure.
Agentic tool calling, retrieval pipelines, prompt versioning and evaluation harnesses — built with the audit trails and authorization boundaries that regulated clients actually require.
Plant engineer and digital transformation analyst at a 1,000 tn/day polymer producer before moving into AI. I have stood on the factory floor the dashboards describe.
MSc in Data Science from the University of Verona, graduated 110 e lode with a scholarship, on top of an MEng in Mechatronics & Robotics with a 4.0 GPA.
Educated and employed across Türkiye, Poland and Italy. Now based in Adana and working remotely for an Italian consultancy, delivering systems for multi-site European clients — including trilingual (EN/IT/DE) production interfaces.
Co-author on two 2025 Chemical Engineering Transactions papers applying neural networks, Bayesian optimization and uncertainty quantification to azeotropic distillation in polymer recycling — plus a thesis on rehabilitation robotics.
See publicationsLooking for someone who can take an AI system from architecture to a deployed, audited, on-premise product? I am open to conversations about senior AI engineering, technical leadership and consulting work. I work remotely from Türkiye and am used to distributed teams across European time zones.