Analyst · Data & AI Ethics @ Deloitte Advisory
ISO/IEC 42001 Lead Auditor · Cybersecurity Specialist
$ whoami
────────────────────────────────────────────────────────────────────────────
simone@deloitte ~
LOCATION Turin, Italy
ROLE Analyst · Data & AI Ethics @ Deloitte Advisory (Milan)
CERTS ISO/IEC 42001 Lead Auditor · Cybersecurity Specialist (EPICODE)
FOCUS AI Governance · Agentic AI Risk · Cybersecurity
INTERESTS LLM red-teaming · agentic-AI threat modeling · ML supply chain
EDU MSc Astrophysics & Cosmology, University of Bologna
NOW AI Solutions Architect master · ProfessionAI
▶ current module: Agentic AI
────────────────────────────────────────────────────────────────────────────
$ _- Day job. AI risk assessment and trustworthy-AI implementation for Deloitte clients (EU AI Act, ISO/IEC 42001, NIST AI RMF). Cyber side: ISO 27701 assessments, policy drafting, awareness campaigns.
- Background. MSc Astrophysics & Cosmology, University of Bologna — thesis on automatic differentiation for gravitational lensing (PyTorch / TensorFlow).
- Publication. Co-author of an AI Human Rights Impact Assessment paper, LSAI 2025.
- Currently studying. AI Solutions Architect master at ProfessionAI —
on the Agentic AI module right now (see
master.logbelow).
focus/
├── ai-engineering/ · building this skill (master + side work)
│ ├── llm-applications · LangChain · prompt pipelines · output parsers
│ ├── rag-systems · ChromaDB · BM25 hybrid retrieval · chunking
│ ├── agentic-ai · multi-agent orchestration · tool use
│ ├── ml-pipelines · feature engineering · model selection
│ └── solution-architecture · cloud AI services · scalable deployment
│
├── ai-governance/ · day job @ Deloitte
│ ├── risk-management · tooling & agentic-AI risk framework
│ ├── compliance · EU AI Act · ISO/IEC 42001 · NIST AI RMF
│ ├── trustworthy-ai · fairness · transparency · robustness
│ └── human-rights-iA · LSAI 2025 paper
│
├── cybersecurity/ · day job @ Deloitte
│ ├── policy-development · drafting & reviewing (public + private)
│ ├── assessments · ISO 27701 · access audits · gap analyses
│ ├── awareness · phishing campaigns · cookie consent (OneTrust)
│ └── frameworks · ISO 27001 · NIST CSF · NIS2
│
└── ai-security/ · research interest
├── llm-red-teaming · prompt injection · jailbreaks
├── agentic-systems · tool-use abuse · multi-agent failures
├── ml-pipeline-threats · supply chain · model integrity
└── hardening · defense-in-depth for AI architectures
ml core ────── scikit-learn · NumPy · Pandas · SciPy
regression · classification · clustering · model selection
cross-validation · feature engineering · pipelines
deep learning PyTorch · TensorFlow · Keras · Pyro (probabilistic)
CNNs · MLPs · automatic differentiation
nlp ────────── HuggingFace Transformers · sentence-transformers
NLTK · Gensim (Word2Vec) · spaCy
tokenization · embeddings · NER · text classification
llm / rag ──── LangChain (core · text-splitters · runnables · parsers)
Ollama (local LLMs) · ChromaDB (vector store)
BM25 (rank_bm25) · hybrid retrieval · prompt engineering
chunking strategies · output parsing · RAG evaluation
agentic ai ─── multi-agent orchestration · tool use · planner/executor
(in study via the current master module)
mlops / cloud Solution architecture · CI/CD for ML · Containers (Docker)
Cloud AI services (AWS · GCP · Azure) · model serving
scalable AI systems · monitoring (in progress)
viz ────────── Matplotlib · Seaborn · Plotly
data ───────── SQL (MySQL · MariaDB) · MongoDB · CSV / JSON pipelines
large-scale astro datasets (former research)
languages ──── Python · Lua · C · Bash · Fortran · SQL
standards ──── EU AI Act · ISO/IEC 42001 · NIST AI RMF
ISO 27001 · ISO 27701 · NIST CSF · NIS2
cybersecurity network · Wireshark · TCPdump · pfSense
pentest · Nmap · Metasploit · Burp Suite · SQLmap · Nessus
malware · IDA Pro · OllyDbg · Splunk
practice · HackTheBox · TryHackMe
systems ────── Windows · Linux · macOS · virtualized environments
Dec 2024 – now Analyst · Data & AI Ethics
Deloitte Advisory · Milan
├─ AI risk management tooling & agentic-AI risk framework
├─ trustworthy-AI guidelines for international public sector
├─ ISO 27701 assessments · cyber policy · access audits
└─ phishing campaigns · awareness · cookie consent (OneTrust)
Nov 2022 – 2023 Research Intern · ML for Astrophysics
INAF / OAS · Bologna
├─ ML models (PyTorch, TensorFlow) for gravitational lensing
└─ deep learning to accelerate large-scale astro pipelines
2018 – present STEM Tutor · Mathematics & Physics
Freelance + schools (Modena · Bologna)
AI_solutions_architect_course —
hands-on projects per module. Public repo, updated as I progress.
✓ done · ▶ in progress · ◌ planned
────────────────────────────────────────────────────────────────────────────
✓ Python Programming warehouse order monitoring system
[Python · stdlib · dataclasses · pathlib]
✓ Machine Learning defective-parts classification pipeline
[scikit-learn · NumPy · Pandas · Matplotlib
· Seaborn · MLP · RF · SVM · KNN · PCA]
✓ Large Language Models RAG for company knowledge base
[LangChain · Ollama · ChromaDB · BM25
· sentence-transformers · HF Transformers
· NLTK · Gensim · hybrid retrieval]
▶ Agentic AI multi-agent orchestration (in progress)
[planner/executor agents · tool use
· agent frameworks]
◌ Business Case & AI PM vision → deployment of an AI system
◌ Cloud for AI multi-cloud AI infrastructure (AWS · GCP · Azure)
◌ AI Service Deployment delivery-time prediction in production
[Docker · CI/CD · model serving · monitoring]
◌ Data Governance & KM governance framework for a pharma company
◌ Solution Design scalable architecture for AI systems
◌ AI Architecture Security risk & vulnerability analysis of an AI system
◌ Governance & Compliance AI governance framework for a fintech
────────────────────────────────────────────────────────────────────────────
Direct links to the completed work:
00_python_programming— order monitoring system (pure Python)01_machine_learning— classification pipeline + 7 exercises (preprocessing, regression, regularization, classification, clustering, NNs, face recognition)02_large_language_models— hybrid RAG (vector + BM25) + 4 exercises (text analysis, NER with BERT, LangChain pipelines, chatbot)03_agentic_ai— multi-agent orchestration (in progress)
2025 – now AI Solutions Architect (Master) · ProfessionAI
2024 Cybersecurity Specialist · EPICODE
2018 – 2024 MSc Astrophysics & Cosmology · University of Bologna
Thesis: Applications of Automatic Differentiation
in Gravitational Lensing (PyTorch · TensorFlow)
Supervisors: Prof. L. Moscardini · Dr. M. Meneghetti
2014 – 2018 BSc Astronomy · University of Bologna
Thesis: Formazione delle Galassie
Supervisor: Prof. D. Dallacasa
- AI Human Rights Impact Assessment — co-authored, LSAI 2025.
editor Neovim 0.11 · lazy.nvim · LSP · Treesitter · Copilot
shell zsh · Powerlevel10k · autosuggestions · syntax-hl
terminal WezTerm / Ghostty · Catppuccin Macchiato · JetBrainsMono NF
search fzf · ripgrep · zoxide · bat
git lazygit · gh CLI
ai Claude Code · GitHub Copilot · local LLMs via Ollama
notes Obsidian
runtime Python (conda) · Lua · Go · C
italian ████████████████ native
english ██████████████░░ C1
spanish ██████████████░░ C1
$ exit
logout