fv . Let's talk

AI/ML DEVELOPER & DATA SCIENTIST

Filippo
Vicentini .

Applied AI for real
industrial and business problems.

Computer Vision Generative AI Predictive ML

I build and deploy machine learning solutions across Computer Vision, predictive modeling, and Generative AI—combining an applied mathematics foundation with hands-on industrial development.

M.Sc. Data Science B.Sc. Applied Mathematics
Explore projects Download CV
WHAT I BUILD
Computer Vision Predictive ML Generative AI Deep Learning Object Detection RAG Systems AI Agents Data Analytics Model Explainability MLOps API Deployment GPU Training

01 / SELECTED PROJECTS

Theory meets
real-world problems.

Four featured case studies across industry,
machine learning, computer vision, and probability.

FEATURED CASE STUDIES 04 projects
Predicted versus actual industrial processing times R² .932
INDUSTRIAL ML

Processing Times

More info Private code
Anonymized industrial technical drawing with YOLO detections YOLO · 3 classes
COMPUTER VISION

Technical Drawing Detection

More info Private code
Predicted versus actual industrial processing times on the validation set
R² 0.932 · MAPE 8.0%
MACHINE LEARNING / INDUSTRY THESIS + WORK

Predict processing times. Improve production.

Prediction of machine processing times from structural and operational features, from model comparison and explainability through to deployment on a Linux server.

  • Python
  • FastAPI
  • Docker
  • Random Forest
  • SHAP
University of Verona × Manni Sipre More info
Anonymized industrial technical drawing with YOLO detections for holes and slots
HOLES · SLOTS · INCLINED CUTS
COMPUTER VISION / INDUSTRY WORK

Object detection in industrial technical drawings.

A YOLO-based workflow that turns technical drawings into machine-readable information by detecting production-relevant geometric features.

  • YOLO
  • Computer Vision
  • Python
  • Object Detection
  • GPU Training
Manni Sipre · Work project More info
ROC curve, precision recall curve, confusion matrix and residual distribution for rain prediction
ROC AUC 0.957
CLASSIFICATION / WEATHER UNIVERSITY

Rain prediction under class imbalance.

A comparative classification study focused on model choice, resampling strategies, and the precision–recall trade-off for next-day rain prediction.

  • Scikit-learn
  • XGBoost
  • Neural Networks
  • SMOTE
More info View on GitHub
Comparison of single and parallel random walk cover-time scaling
n log n → log n
PROBABILITY / SIMULATION UNIVERSITY

Single vs. parallel random walks.

Simulation of graph cover times showing how parallel random walks can dramatically change the scaling behavior of exploration.

  • Python
  • NetworkX
  • Probability
  • Simulation
More info View on GitHub
MORE PROJECTS View all repositories on GitHub
MORE PROJECTS

Additional academic work across data analysis and statistical learning.

DATA ANALYSIS / E-COMMERCE UNIVERSITY

Amazon Sales Analysis

Exploratory analysis of Amazon India sales data, predictive analysis of rejected orders, and interactive visualization of insights.

  • Pandas
  • Python
  • Streamlit
More info View on GitHub
ANOMALY DETECTION / ROBOTICS UNIVERSITY

Anomaly Detection in Robotics

Detection of anomalies in robotic movements using supervised and unsupervised methods, with evaluation of predictive performance and efficiency.

  • Machine Learning
  • Statistical Modeling
More info View on GitHub

02 / EXPERIENCE

From data science.
To production AI.

SEP 2025 — PRESENT

CURRENT

AI/ML Developer

Manni Sipre S.p.A. · Manni Group

Computer Vision Generative AI Deployment Analytics

Building applied AI solutions across computer vision, generative AI, analytics, and software deployment, with a focus on tools that can support real production and business workflows.

Computer Vision

YOLO and OpenCV pipelines for industrial technical drawings, including GPU training, evaluation, and optimization.

Generative AI

RAG prototypes and agentic workflows for documents, internal knowledge, and information retrieval.

Deployment

Applications and microservices built with FastAPI, Docker, and REST APIs.

Analytics

KPI dashboards and data-driven tools designed to support operational visibility and business processes.

SELECTED WORK PROJECT Object Detection in Industrial Technical Drawings

OCT 2024 — AUG 2025

INTERNSHIP

Junior Data Scientist

Manni Sipre S.p.A.

Predictive Modeling Analytics Data Products

Applied machine learning and analytics to industrial and commercial data, working on predictive models and interactive dashboards for operational use cases.

Predictive Modeling

Production-time prediction, customer segmentation, and sales-volume forecasting.

Data Products

Interactive dashboards for KPI monitoring and operational performance analysis.

SELECTED CASE STUDY Predicting Industrial Processing Times

03 / SKILLS

Built around AI.
Grounded in data.

A practical stack for building, evaluating,
and deploying AI solutions end to end.

01MODELS

Machine Learning

Predict · evaluate · explain

XGBoostScikit-learnSHAP
Processing Times
03GEN AI

Generative AI

Retrieve · reason · automate

RAGAgentsLLM workflows
04DEEP

Deep Learning

Train · optimize · generalize

PyTorchNeural NetsGPU Training
05SHIP

Deployment

Serve · containerize · integrate

FastAPIDockerREST APIs
06DATA

Data & Analytics

Explore · visualize · decide

SQLPandasNumPyPower BI
FOUNDATION Python/Applied Mathematics/Statistical Learning
CORE FOUNDATION Python / Applied Mathematics / Statistical Learning
01MODELS

Machine Learning

Predictive modeling, model evaluation and explainability for structured data and real operational problems.

  • XGBoost
  • Scikit-learn
  • SHAP
  • MLflow
02VISION

Computer Vision

Object detection, classification and inspection pipelines applied to industrial technical drawings and visual workflows.

  • Ultralytics YOLO
  • OpenCV
  • PyTorch
  • CUDA
03GEN AI

Generative AI

RAG systems, AI agents and document understanding for retrieval, reasoning and application-oriented workflows.

  • RAG
  • LangChain
  • LlamaIndex
  • Hugging Face
  • AI Agents
04DEEP

Deep Learning

Neural-network training, evaluation and optimization, including GPU-accelerated experimentation for predictive and vision tasks.

  • PyTorch
  • TensorFlow
  • Neural Networks
  • GPU Training
05SHIP

Deployment & MLOps

Turning experiments into usable services and applications with reproducible deployment and integration workflows.

  • FastAPI
  • Docker
  • REST APIs
  • Git
  • Linux
  • MLflow
06DATA

Data & Analytics

Data manipulation, exploration, visualization and reporting to support modeling and operational decision-making.

  • SQL
  • Pandas
  • NumPy
  • Power BI
  • Matplotlib
ADDITIONAL TOOLS

Java · MATLAB · Excel · VS Code

LANGUAGES

Italian — native · English — professional

04 / EDUCATION & CERTIFICATIONS

Academic foundations.
Applied AI learning.

A background in applied mathematics and data science, complemented by focused training in AI agents and retrieval-augmented generation.

2020 — 2023 102/110

B.Sc. Applied Mathematics

University of Verona

CERTIFICATIONS IBM · JUL 2026 Fundamentals of Building AI Agents IBM · AUG 2026 Build RAG Applications: Get Started
01
OCT 2023 — JUL 2025 UNIVERSITY OF VERONA

M.Sc. in Data Science

Focused on data analysis, predictive modeling, Machine Learning and Deep Learning, with practical work in Python and SQL.

THESIS

Predicting industrial machinery processing times using Machine Learning, developed during an internship at Manni Sipre.

View thesis case study
FINAL GRADE 110/110 with honors
02
OCT 2020 — OCT 2023 UNIVERSITY OF VERONA

B.Sc. in Applied Mathematics

Built a strong foundation in algebra, analysis, geometry, computational mathematics, probability and statistics, with applications to financial economics.

THESIS

Simulation of the Credit Default Swap financial instrument using the Java programming language.

FINAL GRADE 102/110