Portfolio
Download PDF

Simone Perini

AI & Software Engineer

Padova, Italy linkedin.com/in/simone-perini-964a54157 github.com/tzii tzii.github.io

AI & Data graduate (EQF 5) with experience at InfoCamere and Galileo Network. I build Python backends for LLM and document-intelligence applications, with vector retrieval, answers linked to their sources and reproducible evaluation. I also maintain two open-source apps with public releases, tests and CI.

Experience

AI & Data Intern

Mar 2026 – May 2026

InfoCamere, Padova

  • Designed and built italianRAG, a FastAPI workspace for document intelligence on Italian corporate PDFs, with typed extraction, facts in SQLite, Qdrant retrieval and answers linked to their evidence.
  • Created a reproducible benchmark of 116 cases. All 116 were answered correctly, supported by citations and checked against a gold standard, within the scope of the deterministic routes.
  • Added cross-document financial and company-register analysis, explicit field states, and optional OCR and vision-model recovery for scanned PDFs.
  • Gathered requirements from colleagues and wrote the documentation for the application.

AI Specialist

Jun 2025 – Sep 2025

Galileo Network

  • Developed and evaluated LLM-based solutions for business processes and document analysis.
  • Compared open-source models, inference stacks and RAG tooling for on-premises deployment.

Projects

ThystTV, open-source Twitch client for Android

github.com/tzii/ThystTV

Kotlin fork of Xtra that I maintain: floating chat, gesture controls, live rewind, foldable layouts and local stats. 16 stable signed releases, more than 3,100 APK downloads, automated tests and CI.

KeeFetch, favicon plugin for KeePass

github.com/tzii/KeeFetch

C# plugin with concurrent downloads, seven icon providers and privacy controls. Its four profiles came out of a benchmark of 19 provider configurations. Unit tests and CI.

Pefforza 4, Connect Four AI

github.com/tzii/Pefforza_4

Python project with a bitboard alpha-beta search, an exact solver, a PPO agent trained with Stable-Baselines3 and OpenCV reading a physical board from a webcam.

ScrapingStore, product-data pipeline

github.com/tzii/ScrapingStore

Playwright scraper for a 3,000-product sandbox store, with run history in SQLite, run-to-run comparisons, static dashboards and CSV export for Power BI.

Education

Higher Technical Diploma in AI & Data

ITS Digital Academy “Mario Volpato”, Padova

2024 – 2026 · EQF level 5

Final grade 107/110

Skills

AI & retrieval

RAG, deterministic extraction, embeddings, reranking, LLM and VLM evaluation, Qdrant

Backend & data

Python, FastAPI, Pydantic, SQL, SQLite, Pandas, NumPy, scikit-learn

Engineering

Git, GitHub, pytest, CI/CD, Docker, benchmarking, technical documentation

Also used

Kotlin/Android, C#/.NET, HTML/CSS/JavaScript, OpenCV, Power BI

Languages

Italian: native

English: C1 (IELTS)

French: B2

I authorise the processing of the personal data in this CV for recruitment purposes, under EU Regulation 2016/679 (GDPR).