ML Systems / AI Infrastructure Engineer
MSc Candidate in Computer Engineering
I’m an ML Systems / AI Infrastructure Engineer based in Italy, with a senior production-grade software engineering background and strong academic foundations in high performance computing, big data and machine learning.
Available for an internship starting in September 2026.
EFashion Software / Alianthis · Nov 2025 – present
Internal project to specialize a code model on a proprietary Java framework. I designed and built the full training pipeline — dataset construction, syntax-aware Java chunking, FIM generation, LoRA-based continued pre-training and fine-tuning, and inference — on the PyTorch ecosystem (Hugging Face Transformers, PEFT), with custom data-packing and training scripts that replaced early Unsloth prototypes.
The dataset covers 40K Java files and reaches ~2.18M CPT examples (~640K L2R, ~1.54M FIM, ~3.8 GB); training runs on rented GPU clusters, with a 3B-parameter model for testing and a 20B target. Preliminary results show clear domain adaptation, measured via held-out perplexity and human evaluation; automated downstream evaluation is next.
EFashion Software / Alianthis · Feb 2026 – present
Production AI platform built around privacy and zero retention by external providers. I designed and operate the entire stack: architecture and deployment automation, LLM gateway integration, document tooling with RAG, search, and transcription, plus observability and pragmatic security engineering.
Built on OpenWebUI, LiteLLM, AWS Bedrock, Langfuse, and Docker with supporting data and monitoring services, it already serves several users, with four models integrated through the gateway — from smaller models to SOTA.
EFashion Software / Alianthis · Apr 2020 – present
Enterprise B2B order platform for fashion, luxury, and sportswear, translated from business requirements and an existing backend product into a new web product. I led the web side from the start: technology choices and React/Relay architecture, GraphQL adoption with API contract design and backend integration, build and delivery workflows with multi-client customization, and coordination of a four-person frontend team.
The platform serves six major luxury, high-fashion, and sportswear brands, with hundreds of users and billions of euros in annual orders managed.
EFashion Software / Alianthis · Jul 2023 – Jan 2024
Frontend foundation and product-platform work for a production quality-control workflow platform, in the same enterprise B2B ecosystem as Order Entry Manager.
EFashion Software / Alianthis · Dec 2021 – Feb 2023
Internal CI/CD and delivery platform on GitHub Actions and AWS, born from the operational complexity of Order Entry Manager. It simplified environment management, supported customer-specific customizations, and made releases repeatable — used for years to automatically spin up hundreds of test environments for new features and to build release candidates.
University of Padua · 2023
BSc thesis in Information Engineering (Automatic coherence evaluation between source code and its related documentation), completed after several years of professional experience. The core is a static analyzer for C90 source code: a custom DFA lexer, a Bison parser, a graph-backed AST with custom data structures, semantic analysis and diagnostics, and manual memory management validated with Valgrind.
University of Padua · Oct 2024 – Jan 2025
Project for the Learning From Networks MSc course: large-scale graph construction, community detection, and graph analytics on the Wikipedia graph — ~3M vertices and ~55M edges. Built in Python with igraph, comparing Leiden, Louvain, and label propagation on modularity and partition stability (adjusted Rand index), plus centrality and inter-cluster analysis.
University of Padua · Apr 2025 – Jun 2025
MSc project for the Artificial Intelligence course: a prototype for institutional canteen planning that combines ranked pairs voting, a genetic algorithm, and multi-objective optimization.
2014 – 2019
Self-taught programming and small software projects, moving into client work from 2017.
EFashion Software / Alianthis · Apr 2020 – Oct 2025
Built and led the Order Entry Manager web platform and its four-person frontend team from the ground up; along the way, created the CI/CD and delivery platform and the frontend foundation of the Quality Control Platform.
University of Padua · 2023
Completed in parallel with full-time work, graduating with full marks; thesis on a static code analyzer.
Jan 2024 – Sep 2024
Parallel freelance engagement: an urban furniture catalog built end-to-end on Next.js, Strapi, and TypeScript — CMS modeling, multilingual content, quote capture, Excel export, and ERP integration.
University of Padua · 2023 – present
High Performance and Big Data Computing curriculum, in parallel with full-time work — with focus on ML, optimization, and graph analytics.
EFashion Software / Alianthis · Nov 2025 – present
Current focus: a domain-specific code LLM post-training pipeline for a proprietary Java framework, and a production zero-retention AI platform for privacy-sensitive deployments.
High Performance and Big Data Computing
University of Padua · 2023 – 2027 (expected)
Average grade: 28+/30
Thesis: Expected spring 2027
University of Padua · 2018 – 2023
Average grade: 28.7/30
Final grade: 110/110
Thesis: C source code static analyzer
I started programming early, driven by a long-standing interest in programming languages, algorithms, data structures, and systems. Professionally, I first built and led enterprise web platforms for B2B workflows in fashion, luxury, and sportswear. I’ve worked across architecture, API contracts, team coordination, customization models, build systems, and delivery.
I’m currently completing my MSc in Computer Engineering, with a curriculum in High Performance and Big Data Computing, at the University of Padua. My recent work has focused on a domain-specific code LLM post-training pipeline and a production zero-retention AI platform. My current work and interests sit at the intersection of ML systems, AI infrastructure, large-scale data processing, performance-oriented computing, and code intelligence.
I’m looking for projects where I can combine production-grade software engineering with my academic background in advanced algorithms, high performance computing, big data, and AI/ML systems.
Outside of work, I’m passionate about ethnomusicology and the traditional music and dances of the Mediterranean, particularly Southern Italy. I’m now studying the region’s frame drum techniques and the related dances. I love traveling to join the traditional contexts where these practices are still alive.