VF.
Software Engineering • Data Engineering • Applied AI

Architecting High-Performance Systems, ETL Pipelines & Applied AI

> Bridging reliable software engineering, scalable data architectures & production machine learning
5+ Years Active Production Exp
10+ Core Engineering Projects
AWS, RabbitMQ & Microservices
Computer Vision & Applied AI
Professional Profile

Three Integrated Engineering Disciplines

Software Engineer

Backend & Distributed Microservices

Architecting resilient backend systems with Python (FastAPI), Node.js, and TypeScript. Specializing in event-driven architectures with RabbitMQ, asynchronous background task dispatching, secure REST/GraphQL APIs, and Dockerized microservices.

Python / FastAPIRabbitMQNode.js / ExpressDockerPostgreSQL

Data Engineer

High-Throughput ETL & Lakehouse Systems

Designing automated extraction, transformation, and load (ETL) pipelines processing multi-source enterprise retail and policy datasets. Certified in Microsoft Fabric Lakehouse architectures, batch/streaming pipelines, relational normalization, and AWS cloud storage.

MS Fabric CertifiedPandas / ETLAWS S3 & AuroraData ModelingGitLab CI/CD

Data Scientist / AI Engineer

Machine Learning, Vision & Production LLMs

Training and deploying production ML systems: YOLOv8 multi-camera computer vision for industrial anomaly detection, Generative AI & Claude/GPT copilot integration with streaming SSE, Qdrant vector databases, and predictive price forecasting (Prophet).

YOLOv8 VisionPyTorchQdrant VectorsLLMs / Prompt Eng.Forecasting
Career Timeline

Professional Work Experience

Software Engineer & System Architect

Live Tech Srl
01/07/2021 – Present

Taking autonomous architectural decisions and end-to-end engineering ownership across the entire software development lifecycle for mission-critical enterprise systems.

Applied AI, LLMs & Production AI Agents: Designed, benchmarked, and deployed enterprise conversational AI agents and LLM pipelines for semantic search, contextual question-answering over corporate knowledge bases, and automated intent classification with sub-second response times.
Industrial Computer Vision & Edge AI: Engineered automated visual defect inspection pipelines combining multi-camera industrial feeds with custom fine-tuned YOLO deep neural networks for real-time quality control.
Distributed Message Brokers & Cloud Architecture: Architected fault-tolerant, event-driven message dispatching systems with RabbitMQ and AWS (S3, Aurora RDS), decoupling asynchronous AI inference and guaranteeing zero-loss transaction routing.
Automated Data Engineering & High-Throughput ETL: Built resilient data processing pipelines handling large retail catalog synchronizations, automated certified email (PEC) categorization, and structured OCR document parsing.
Containerization & MLOps / CI-CD: Maintained Docker containerized microservices, automated regression testing, and CI/CD pipelines across staging and production clusters.
PythonAI Agents & LLMsYOLO VisionFastAPIAWS (S3/Aurora)RabbitMQDockerGitLab CI/CDPostgreSQL
Academic Background

Education & Professional Certifications

2021 – Present

MSc in Computer Engineering: AI & Machine Learning

Roma Tre University

Advanced coursework in Deep Learning, Operations Research & Metaheuristic Optimization, Big Data Analytics, Statistical Learning, and Distributed Computing.

Completed

BSc in Computer Engineering

Università Telematica Internazionale UNINETTUNO

Rigorous foundations in Software Engineering, Algorithms & Data Structures, Relational Database Management, Operating Systems, and Networking.

Microsoft Certified

Microsoft Fabric Data Engineer

Microsoft Official Certification

Lakehouse architecture design, Delta Parquet transformation, Apache Spark data engineering, and scalable enterprise data warehousing.

Industry Badges

Cisco Cybersecurity & Kaggle Data Science

Cisco Networking Academy & Kaggle

Network vulnerability testing, threat mitigation principles, advanced Pandas data manipulation, and exploratory data analysis (EDA).

Engineering Portfolio

Featured Engineering Projects

Divided into three distinct categories: Enterprise Systems, University AI Research, and Personal/Client Platforms.

1. Enterprise & Industrial Systems (Work Experience)

Mission-critical backends, distributed message brokers, and automated data pipelines architected for production environments.

Distributed Architecture Production
Enterprise Asynchronous Message Broker & Dispatcher

Engineered a high-throughput, decoupled event-driven dispatching system leveraging RabbitMQ and Python/FastAPI microservices. Purpose-built to process, validate, and securely route corporate communications (including legal certified PEC emails and automated notifications) with zero data loss, concurrency controls, and comprehensive dead-letter audit logging.

Worker pool architecture with rate-limiting & backoff retry logic Fully decoupled publish-subscribe queues preventing system bottlenecks
RabbitMQPythonFastAPIDockerPostgreSQL
Big Data & Data Engineering Production
High-Throughput Enterprise Retail ETL Pipeline

Automated data extraction, transformation, and loading pipeline designed to synchronize large-scale commercial catalog feeds, dynamic pricing updates, and multi-store inventories. Built with modular Python transformers, automated schema validation, and GitLab CI/CD pipelines to guarantee data integrity across distributed databases.

Automated batch and incremental delta data synchronization Reconciliation algorithms detecting catalog anomalies and discrepancies
PythonPandasGitLab CI/CDDockerETL
Computer Vision & Deep Learning Applied AI
Industrial Multi-Camera Vision & YOLO Defect Inspection

Developed an automated visual inspection and defect detection pipeline integrating synchronized industrial camera streams with custom fine-tuned YOLO (You Only Look Once) deep neural networks. Features synthetic dataset augmentation, custom bounding-box annotation workflows, and real-time inference for automated quality assurance.

Multi-angle camera stream ingestion and synchronized frame processing Optimized inference delivering millisecond defect classification
YOLOv8PyTorchOpenCVCUDAPython
Document AI & Cloud Enterprise Solution
Insurance Policy & DPI Document Processing Suite

Created a microservice pipeline (format extractors, message brokers, and dispatchers) that automatically parses, sanitizes, and normalizes complex insurance contracts, PDF documentation, and DPI safety compliance forms into relational databases with AWS S3 archival.

Modular extractor-broker pattern for zero-touch document digestion High-accuracy regex/OCR field extraction with strict validation schemas
FastAPIAWS S3MySQLDocument AIDocker

2. University & AI Research Projects

Academic research in machine learning, combinatorial optimization, audio neural modeling, and big data analysis.

Big Data & Data Science 5 Notebooks Suite
Yelp Big Data Intelligence & Hybrid Recommender System

Comprehensive end-to-end data science study performed on the Yelp Academic Dataset (150,000+ businesses, millions of user reviews). Structured across 5 modular research notebooks featuring geospatial GIS density maps (Folium), NLP sentiment analysis, TF-IDF topic extraction, and collaborative filtering recommendation models.

Geospatial heatmaps & demographic cluster discovery Matrix factorization & hybrid recommendation algorithms
PythonScikit-LearnNLPFoliumCollaborative Filtering
Operations Research & Algorithms Live Interactive Demo
Combinatorial & Heuristic Scheduling Optimization

Mathematical formulation and algorithmic engine solving NP-Hard job scheduling constraints (maximizing minimum job utility under release dates). Implemented pairwise swap Local Search, Tabu Search, Simulated Annealing, and Iterated Local Search with an interactive Streamlit dashboard.

Implements 4 release-date paper theorems and benchmark evaluations Live interactive execution with customizable algorithm parameters
PythonStreamlitMetaheuristicsSimulated AnnealingTabu SearchOperations Research
Deep Generative AI PyTorch Model
Conditional Variational Autoencoder (CVAE) Audio Synthesis

Designed and trained a deep generative Conditional Variational Autoencoder in PyTorch to synthesize, interpolate, and reconstruct acoustic spectrograms conditioned on target sound features. Implemented continuous latent space arithmetic, STFT audio representations, and Griffin-Lim phase reconstruction.

Continuous smooth traversal and interpolation in acoustic latent space Custom multi-component loss functions (KL divergence + reconstruction loss)
PyTorchLibrosaDeep LearningCVAEAudio DSP
Exploratory Data Analysis Kaggle Analytics
Amazon Sales & Pricing Dynamics EDA

Comprehensive statistical data analysis on 1,400+ Amazon commercial products examining promotional discount mechanics, price elasticity distributions, customer review ratings, and high-frequency keyword sentiment.

Univariate & bivariate distribution modeling with Seaborn Correlation analysis between discount percentage and rating scores
PythonPandasSeabornMatplotlibStatistical EDA

3. Personal & Client Applications (Full-Stack & Mobile)

Real-world production applications, mobile platforms with AI copilots, e-commerce, and enterprise web solutions.

Healthcare Digital Platform Production Client
Sanitaria Asia: Medical Equipment & Orthopedic Portal

Enterprise web presence and digital product catalog for a specialized medical, healthcare, and orthopedic equipment provider. Built with Next.js and React, featuring optimized media delivery, interactive medical supply catalogs, patient inquiry workflows, and Dockerized production deployment.

Optimized Core Web Vitals for instant browsing and catalog search Docker container deployment ensuring dependable high uptime
Next.jsReactDockerResponsive UITailwind CSS
Community Portal & Booking Production Live
AGESCI Guidonia 1 Official Portal

Production web platform live at guidoniauno.it serving the historic AGESCI Guidonia 1 organization. Includes scout chalet & base reservation inquiry systems, dynamic photo archives with lightbox optimization (Sharp), Google APIs integration, and automated email notifications (Nodemailer).

Integrated scout base reservation inquiry engine with Nodemailer Server-side image optimization pipeline utilizing Sharp & Vercel Blob
Next.jsTypeScriptGoogle APIsSharpVercel
Get In Touch

Let's Connect & Build Resilient Systems

I am always interested in discussing software engineering challenges, data pipelines, distributed systems, and cutting-edge applied AI opportunities.

© 2026 Valerio Fiorentini — Software Engineer • Data Engineer • Data Scientist