▸ Blueprint Mode attivo ▸ Blueprint Mode engaged
Santeramo in Colle · IT EST. 1989 Disponibile per progetti selezionati Available for selected projects

Francesco
Perniola

Senior Software EngineerSolution ArchitectAI Engineer

Senior Software Engineer. Progetto sistemi che devono funzionare sotto vincoli reali — CPU-only, on-premise, clienti in produzione, deadline non negoziabili. È in quei vincoli che si vede se un'architettura regge. Senior Software Engineer. I design systems that must work under real-world constraints — CPU-only, on-premise, clients in production, non-negotiable deadlines. It's in those constraints that you see if an architecture holds up.

Esplora il sistema Explore the system

L.04 INTELLIGENCE

L.04 INTELLIGENCE

Feature LLM production-grade — RAG, agenti, chatbot domain-specific — consegnate ai clienti. In parallelo, reti neurali R&D per medical imaging e computer vision industriale.

Production-grade LLM features — RAG, agents, domain-specific chatbots — embedded in client products. In parallel, R&D neural networks for medical imaging and industrial computer vision.

L.03 APPLICATION

Backend NestJS (5+ anni), Django/Celery, API REST e GraphQL. Frontend React + TypeScript + Vite, mobile con React Native per iOS e Android.

NestJS backends (5+ years), Django/Celery, REST and GraphQL APIs. Frontend React + TypeScript + Vite, mobile with React Native for iOS and Android.

L.02 DATA

PostgreSQL + TimescaleDB per time-series, Redis per caching e correlazione event-driven, MQTT per messaging IoT asincrono cross-module.

PostgreSQL + TimescaleDB for time-series, Redis for caching and event-driven correlation, MQTT for asynchronous cross-module IoT messaging.

L.01 INFRASTRUCTURE

Docker e Kubernetes principalmente on-premise. Pipeline CI/CD su GitHub e GitLab. Testing sistematico unit, integration e E2E con Playwright.

Docker and Kubernetes primarily on-premise. CI/CD pipelines on GitHub and GitLab. Systematic unit, integration and E2E testing with Playwright.

§01

Profilo Profile

FULL-STACK · ARCHITECT · AI

Progetto e costruisco sistemi software che vivono in produzione operativa. Guido un team di ~10 ingegneri su prodotti a TRL 6–9 — dai sistemi IoT industriali event-driven alle piattaforme smart city multi-modulo, fino all'integrazione LLM production-grade nei prodotti consegnati ai clienti. L'architettura, per me, non è un artefatto a parte: è la forma che assume il codice quando smette di essere un prototipo.

In parallelo, faccio R&D in deep learning e computer vision — reti neurali per la segmentazione di masse tumorali da scansioni CT, classificazione delle condizioni del manto stradale — con pipeline DICOM scritte da zero in PyTorch e TensorFlow. Il mio centro di gravità è l'infrastruttura on-premise dove software, AI e sistemi distribuiti devono convivere come un'unica cosa, non come moduli separati.

I design and build software systems that live in operational production. I lead a team of ~10 engineers on TRL 6–9 products — from event-driven industrial IoT systems to multi-module smart city platforms, including production-grade LLM integrations shipped inside client products. For me, architecture isn't a separate artifact: it's the shape the code takes when it stops being a prototype.

In parallel, I run R&D in deep learning and computer vision — neural networks for tumor detection and segmentation from CT scans, road surface condition classification — with DICOM pipelines written from scratch in PyTorch and TensorFlow. My center of gravity is the on-premise infrastructure where software, AI, and distributed systems must live as one, not as separate modules.

Ruoli
Roles
Senior SWE · Solution Architect · AI Engineer · Tech Lead
Team
Team
~10 ingegneri · PM · PMT · cliente
~10 engineers · PM · PMT · client
Maturità
Maturity
TRL 6–9 · production
Stile architetturale
Architectural style
Event-driven · DDD · Microservizi
Base
Based in
Santeramo in Colle (BA), Italia
15+
Anni di ingegneria
Years in engineering
10
Ingegneri guidati
Engineers led
6–9
TRL produzione
TRL production
5+
Anni NestJS / Node
Years NestJS / Node
§02

Aree di competenza Areas of expertise

FIVE DOMAINS · FULL DEPTH
E.01 · L.04 INTELLIGENCE

LLM & GenAI in produzione in production

Feature AI-powered consegnate ai clienti, non prototipi. Architettura, governance e integrazione deep nei prodotti.

AI-powered features shipped to clients, not prototypes. Architecture, governance, and deep product integration.

  • Sistemi RAG production-grade con vector store embedded
  • Agenti LLM con tool calling e orchestrazione multi-step
  • Chatbot domain-specific — ogni cliente vuole che "suoni come casa sua", quindi i guardrail li costruisco sul vocabolario e sulle eccezioni del suo dominio, non su liste generiche
  • Copilot interni come acceleratori di sviluppo
  • Prompt engineering per sviluppo, controllo e governance
  • Integrazione LLM embedded in piattaforme on-premise
  • Production-grade RAG systems with embedded vector stores
  • LLM agents with tool calling and multi-step orchestration
  • Domain-specific chatbots — each client wants theirs to "sound like home", so I build guardrails on their vocabulary and edge cases, not on generic lists
  • Internal copilots as development accelerators
  • Prompt engineering for development, control, governance
  • Embedded LLM integration in on-premise platforms
Scope · Production · client-delivered
Scope · Production · client-delivered
E.02 · L.04 INTELLIGENCE

Deep Learning & Computer Vision & Computer Vision

Reti neurali per medical imaging e monitoraggio di infrastrutture fisiche. Pipeline custom, niente MONAI.

Neural networks for medical imaging and physical-infrastructure monitoring. Custom pipelines, no MONAI.

  • Tumor detection & segmentation da scansioni CT
  • Classificazione delle condizioni del manto stradale
  • Pipeline DICOM custom end-to-end
  • Training su hardware dedicato, inference on-premise
  • PyTorch e TensorFlow (entrambi in produzione di ricerca)
  • Integrazione dei modelli direttamente nelle piattaforme target
  • Tumor detection & segmentation from CT scans
  • Road surface condition classification
  • Custom end-to-end DICOM pipelines
  • Training on dedicated hardware, on-premise inference
  • PyTorch and TensorFlow (both in research production)
  • Model integration directly into target platforms
Scope · R&D · on-premise deploy
Scope · R&D · on-premise deploy
E.03 · § LEADERSHIP

Solution Architecture & Tech Lead Architecture & Tech Lead

Dal requisito al deploy. Guida tecnica di un team di ~10 persone e interfaccia diretta con stakeholder business.

From requirement to deploy. Technical leadership of a ~10-person team and direct interface with business stakeholders.

  • Analisi dei requisiti e selezione dello stack
  • High-Level Design e Low-Level Design
  • Progettazione algoritmica e report di sviluppo
  • Quando il cliente deve scegliere tra alternative, non gli do una raccomandazione sola: preparo un confronto con Pro, Contro e "quando è la scelta giusta" per ciascuna opzione, incluso il costo totale nel tempo
  • Stima effort (G/U) su team multi-ruolo
  • Gestione di ~10 sviluppatori · code review · mentoring
  • Interfaccia con PM, Project Management Team, clienti
  • Requirements analysis and stack selection
  • High-Level Design and Low-Level Design
  • Algorithm design and development reports
  • When the client has to choose between alternatives, I don't hand over a single recommendation: I prepare a comparison with Pros, Cons, and "when it's the right fit" for each option, including total cost of ownership
  • Effort estimation (D/U) across multi-role teams
  • Leading ~10 engineers · code review · mentoring
  • Interface with PM, Project Management Team, clients
Scope · end-to-end · client-facing
Scope · end-to-end · client-facing
E.04 · L.04 INTELLIGENCE

Applied ML production-grade production-grade

ML pragmatico che gira in produzione reale — con vincoli CPU-only e requisiti di latenza bassi.

Pragmatic ML running in real production — with CPU-only constraints and tight latency requirements.

  • Anomaly detection statistica multi-sensore
  • Baseline learning passivo con calibrazione continua
  • XGBoost per forecasting e classificazione
  • Regressione lineare per stime real-time
  • Algoritmi predittivi su dati IoT time-series
  • Triangolazione logaritmica e correlazione asincrona
  • Multi-sensor statistical anomaly detection
  • Passive baseline learning with continuous calibration
  • XGBoost for forecasting and classification
  • Linear regression for real-time estimation
  • Predictive algorithms over IoT time-series data
  • Logarithmic triangulation and asynchronous correlation
Scope · Production · IoT + smart city
Scope · Production · IoT + smart city
E.05 · § GOVERNANCE

Risk Management & Migration Safety & Migration Safety

Disciplina formale di risk & estimation governance. Ogni decisione tecnica client-facing è accompagnata da un framework che espone rischi, assunzioni e varianza — non da promesse.

Formal risk & estimation governance discipline. Every client-facing technical decision comes with a framework that exposes risks, assumptions, and variance — not promises.

  • Risk register formale con matrice Impatto × Probabilità × Mitigazione
  • Stime con varianza esplicita e forchetta motivata (min–max)
  • Assunzioni documentate con trigger espliciti di re-scope — se un'assunzione decade, il cliente sa subito che la stima non tiene più e si rinegozia, non si scopre in ritardo
  • Migration safety patterns: Adapter layer, Shadow testing, contract preservation
  • Gap analysis pre-implementazione con fallback logics documentate
  • Articolazione in cantieri separati (stabilità/refactoring · conformità · evolutive)
  • Formal risk register with Impact × Probability × Mitigation matrix
  • Estimates with explicit variance and justified range (min–max)
  • Documented assumptions with explicit re-scope triggers — if an assumption breaks, the client knows immediately that the estimate no longer holds and we renegotiate, instead of discovering it too late
  • Migration safety patterns: Adapter layer, Shadow testing, contract preservation
  • Pre-implementation gap analysis with documented fallback logics
  • Work stream partitioning (stability/refactoring · compliance · evolutions)
Scope · client-facing assessment · pre-implementation
Scope · client-facing assessment · pre-implementation
§03

Opere significative Signature works

ARCHETIPI · SOTTO NDA ARCHETYPES · UNDER NDA
A.01 2023 —

Piattaforma Smart City & Mobility Smart City & Mobility Platform

SOLUTION ARCHITECT · TECH LEAD

MVP con un vincolo che ha guidato ogni scelta tecnica: deve partire su una macchina di sviluppo da 16 GB senza GPU, con un solo docker-compose. Ho scartato LSTM e Random Forest in favore di XGBoost su CPU, scelto Mosquitto come bus di comunicazione tra moduli invece di RabbitMQ per tenere leggera l'infrastruttura, e tenuto Python solo per la parte ML — tutto il resto è TypeScript, così il team non cambia linguaggio a metà giornata.

An MVP built around one constraint that drove every technical choice: it had to run on a 16 GB dev machine with no GPU, via a single docker-compose. I dropped LSTM and Random Forest in favor of CPU-bound XGBoost, picked Mosquitto over RabbitMQ as the inter-module bus to keep the infrastructure light, and kept Python only for the ML layer — everything else is TypeScript, so the team doesn't switch languages mid-day.

STACK
NestJSReact · VitePostgreSQL TimescaleDBRedisMQTT (Mosquitto) XGBoostDocker Compose
A.02 2024 —

Sistema IoT per Sicurezza Industriale Industrial Safety IoT System

SOLUTION ARCHITECT

Il cliente voleva dimezzare il numero di sensori su strutture già installate senza perdere copertura di monitoraggio. La sfida: zero modifiche al firmware, zero costi hardware aggiuntivi. Ho spostato l'intelligenza nel cloud — finestra scorrevole per correlare messaggi asincroni, voting spaziale per distinguere urti reali da vibrazioni accidentali, triangolazione per localizzare l'impatto anche su montanti ora privi di sensore diretto.

The client wanted to halve the sensor count on already-installed structures without losing monitoring coverage. The constraint: zero firmware changes, zero additional hardware cost. I moved the intelligence to the cloud — sliding window to correlate asynchronous messages, spatial voting to tell real impacts from accidental vibrations, triangulation to locate impacts even on struts now missing a direct sensor.

STACK
PythonDjangoCelery RedisPostgreSQL Event-drivenDigital Twin
A.03 2022 —

Feature LLM Production nei prodotti cliente Production LLM features in client products

AI ENGINEER · ARCHITECT

Sistemi RAG, agenti con tool calling, chatbot domain-specific e copilot interni integrati dentro prodotti software consegnati ai clienti. Prompt engineering per sviluppo, controllo e governance di AI in pipeline reali.

RAG systems, agents with tool calling, domain-specific chatbots, and internal copilots embedded inside software products delivered to clients. Prompt engineering for development, control, and governance of AI in real pipelines.

APPROCCIO APPROACH
RAGLLM Agents API-native SDK API-native SDK Embedded deploy Embedded deploy Prompt Engineering
A.04 R&D

Medical Imaging — Tumor Segmentation Medical Imaging — Tumor Segmentation

ML RESEARCH ENGINEER

Reti neurali per detection e segmentazione di masse tumorali da scansioni CT. Pipeline DICOM scritte da zero, training su hardware dedicato, inferenza integrata nella piattaforma target on-premise. Parallelamente: classificazione del manto stradale.

Neural networks for tumor detection and segmentation from CT scans. DICOM pipelines written from scratch, training on dedicated hardware, inference embedded in the target on-premise platform. In parallel: road surface classification.

STACK
PyTorchTensorFlow DICOMComputer Vision Training GPU dedicato Dedicated GPU training
A.06 2024 —

Piattaforma Retail Digital Twin Retail Digital Twin Platform

SOLUTION ARCHITECT · TECH LEAD

Il sistema nasceva shelf-centric — "lo scaffale è vuoto, riempilo". Ma in GDO questo non basta: serve sapere se un prodotto sta invadendo lo spazio di un altro, se la posizione effettiva corrisponde al planogramma. Ho portato il modello a product-shelf-centric, con allocazioni per GTIN/EAN, mismatch detection e OSA calcolato a livello di singola referenza. Multi-tenancy con Row-Level Security perché il cliente finale gestisce store di insegne diverse.

The system started shelf-centric — "the shelf is empty, refill it". But in large-scale retail that's not enough: you need to know if a product is invading another's space, if actual positioning matches the planogram. I moved the model to product-shelf-centric, with per-GTIN/EAN allocations, mismatch detection, and OSA computed at single-SKU granularity. Multi-tenancy with Row-Level Security because the end client runs stores across different banners.

STACK
DjangoDRF MQTT (enterprise)Celery RedisPostgreSQL + TimescaleDB Apache Superset
A.07 2025 —

Piattaforma clinica — Gap analysis & roadmap Clinical Data Platform — Gap analysis & roadmap

SOLUTION ARCHITECT · ANALYSIS LEAD

La piattaforma calcolava età biologica e capacità vitale, ma gli stessi valori comparivano in punti diversi con logiche divergenti e non si poteva ricostruire con certezza quale versione di calcolo avesse prodotto un referto. Ho prodotto un assessment completo che il cliente potesse leggere con il proprio board: 299 gg/uomo stimati, articolati in tre cantieri separati — refactor di stabilità, allineamento ai requisiti documentali, evolutive ML e cluster analysis. Ogni cantiere può partire senza bloccare gli altri.

The platform computed biological age and vital capacity, but the same values appeared in different places with diverging logic, and you couldn't reliably reconstruct which calculation version had produced a report. I delivered a complete assessment the client could walk through with their board: 299 estimated person-days, partitioned into three independent work streams — stability refactor, alignment with documented requirements, ML evolutions and cluster analysis. Each stream can start without blocking the others.

AMBITO SCOPE
Gap AnalysisWBS Stima gg/uomo Person-day estimation Refactoring roadmap Refactoring roadmap Standardizzazione dati clinici Clinical data standardization Audit trail
A.08 2026 —

Integrazione Enterprise SaaS Enterprise SaaS Integration

SOLUTION ARCHITECT · INTEGRATION LEAD

Adapter layer tra gestionale cliente e piattaforma e-learning enterprise. Matrice RACI su team cross-azienda, REST API con token security e permessi granulari, data seeding per allineamento utenze pregresse. Pattern di integrazione riutilizzabile su altri provider.

Adapter layer between the client's business-management system and an enterprise e-learning platform. RACI matrix across cross-company teams, REST APIs with token security and granular permissions, data seeding for legacy-user alignment. Reusable integration pattern across providers.

AMBITO SCOPE
REST APIsToken Security Matrice RACI RACI matrix Data Seeding Integrazione sistemi esterni External system integration WBS unificata Unified WBS
A.05 2026 —

Linee Guida Operative per lo Sviluppo Software Software Development Operating Guidelines

METHODOLOGY ARCHITECT · AI GOVERNANCE

Formalizzazione dell'SDLC aziendale: matrice di responsabilità a quattro ruoli (Analisi, ML, Sistemi, Sviluppo), tre macro-workflow (Kanban → Scrum, Water-Scrum-Fall per legacy, Day-2 operations), governance esplicita degli agenti AI mappati alle fasi di sviluppo, Knowledge Management via RAG aziendale su tutti i Report di Esecuzione.

Formalization of the company SDLC: four-role responsibility matrix (Analysis, ML, Systems, Development), three macro-workflows (Kanban → Scrum, Water-Scrum-Fall for legacy refactoring, Day-2 operations), explicit governance of AI agents mapped to development phases, Knowledge Management via corporate RAG over all Execution Reports.

AMBITO SCOPE
Responsibility Matrix Responsibility Matrix SDLC AI Governance RAG aziendale Corporate RAG LlamaIndexPinecone HLD / LLD
§04

Esperienza Experience

CHRONOLOGICAL
2021 — presente— present Dyrecta Lab · Conversano (BA)

Senior Software Engineer · Solution Architect

TECH LEAD · ~10 ENGINEERS · METHODOLOGY · CLIENT-FACING
  • Guida tecnica di ~10 sviluppatori, coordinamento con Project Management Team, Project Manager e clienti finali.
  • Autore delle Linee Guida Operative per lo Sviluppo Software aziendali: matrice di responsabilità a quattro ruoli, macro-workflow MVP → produzione, Water-Scrum-Fall per legacy, Day-2 operations, governance degli agenti AI nell'SDLC e Knowledge Management via RAG aziendale.
  • Nominato all'unanimità dall'esecutivo come Referente Analisi e di Area (RA) ad interim: ingegneria dei requisiti, conformità normativa, QA tecnica, gestione rischi e consegna finale su tutti i progetti attivi, in parallelo al ruolo primario di Referente Sviluppo Software (RD).
  • Assessment client-facing con risk register formale (Impatto × Probabilità × Mitigazione), stime con varianza esplicita e assunzioni documentate, WBS per tempo probabile articolata in cantieri separati (stabilità, conformità, evolutive), sintesi manageriali con impact business.
  • Architettura software end-to-end su prodotti TRL 6–9: analisi dei requisiti, selezione stack, High-Level Design, Low-Level Design, progettazione algoritmica, report di sviluppo.
  • Progettazione di piattaforme IoT industriali event-driven con MQTT, correlazione multi-sensore su timestamp asincroni, triangolazione logaritmica, digital twin geometrico, anomaly detection statistica con baseline learning passivo.
  • Architettura e sviluppo di piattaforma smart city multi-modulo (air quality, sharing, parking, traffic monitoring) con MQTT cross-module, Docker Compose, PostgreSQL + TimescaleDB, Redis.
  • Delivery di feature AI-powered production-grade in prodotti cliente: sistemi RAG, agenti, chatbot domain-specific, copilot interni; prompt engineering per sviluppo, controllo e governance di AI.
  • R&D in deep learning e computer vision: reti neurali per segmentazione di masse tumorali da scansioni CT e classificazione delle condizioni del manto stradale, con pipeline DICOM custom (PyTorch, TensorFlow).
  • Technical leadership of ~10 engineers; coordination with Project Management Team, Project Managers, and client stakeholders.
  • Author of the company's Software Development Operating Guidelines: four-role responsibility matrix, MVP → production macro-workflow, Water-Scrum-Fall for legacy refactoring, Day-2 operations, AI agent governance across the SDLC, and Knowledge Management via corporate RAG.
  • Unanimously appointed by the executive team as interim Analysis & Area Lead (RA): requirements engineering, regulatory compliance, technical QA, risk management, and final delivery across all active projects, in parallel with the primary Software Development Lead (RD) role.
  • Client-facing assessments with formal risk register (Impact × Probability × Mitigation), estimates with explicit variance and documented assumptions, WBS by probable time partitioned into separate work streams (stability, compliance, evolutions), managerial synthesis with business impact.
  • End-to-end software architecture on TRL 6–9 products: requirements analysis, stack selection, High-Level Design, Low-Level Design, algorithm design, development reports.
  • Design of event-driven industrial IoT platforms with MQTT, multi-sensor asynchronous timestamp correlation, logarithmic triangulation, geometric digital twins, and statistical anomaly detection with passive baseline learning.
  • Architecture and development of a multi-module smart-city platform (air quality, sharing, parking, traffic monitoring) with cross-module MQTT, Docker Compose, PostgreSQL + TimescaleDB, Redis.
  • Delivery of production-grade AI-powered features in client products: RAG systems, agents, domain-specific chatbots, internal copilots; prompt engineering for development, control, and governance of AI.
  • R&D in deep learning and computer vision: neural networks for tumor detection and segmentation from CT scans, and road surface condition classification, with custom DICOM pipelines (PyTorch, TensorFlow).
2017 — 2020 Evolware srl · Gioia del Colle (BA)

Software Developer · Team Lead

WEB APPLICATIONS · ENTERPRISE INTEGRATIONS
  • Coordinamento del team di sviluppo e sviluppo full-stack di applicativi web.
  • Gestione del ciclo di vita di progetti web con integrazioni di sistema e backend enterprise.
  • Coordination of the development team and full-stack development of web applications.
  • Lifecycle management of web projects with system integrations and enterprise backends.
§05

Stack tecnologico Technology stack

DEPTH · IN ANNI DEPTH · IN YEARS
Backend
Node.js5+ NestJS5+ TypeScript5+ Python10+ Django Celery PHP10+ REST GraphQL SOAP WebSocket MQTT RabbitMQ
Frontend & Mobile
React10+ TypeScript5+ Vite React Native8+ iOS Android HTML515+ CSS315+
Data & Persistence
PostgreSQL TimescaleDB MySQL15+ Redis Firestore SQL15+ SQLAlchemy
Infrastructure / DevOps
Docker Docker Compose Kubernetes Apache Nginx IIS Firebase GitHub GitLab CI/CD
AI / ML
PyTorch TensorFlow XGBoost DICOM Computer Vision RAG LLM Agents Prompt Engineering
Architettura Architecture
Event-driven Domain-Driven Design Microservizi Microservices Digital Twin HLD / LLD
Qualità & Testing Quality & Testing
Unit Integration E2E Playwright Cypress Jest Vitest pytest Code Review GDPR Industry 4.0
§06

Formazione & lingue Education & languages

FOUNDATIONS
2021 — 2022

Laurea Triennale in Ingegneria Informatica (L-8) Bachelor's Degree in Computer Engineering (L-8)

Universitas Mercatorum, Roma — conseguita in circa un anno e mezzo in parallelo al lavoro full-time.
Universitas Mercatorum, Rome — earned in roughly one and a half years while working full-time.
2004 — 2009

Diploma di Maturità Scientifica PNI Scientific Secondary School Diploma (PNI)

Liceo Scientifico R. Canudo · Gioia del Colle
Italiano madrelingua
Italian native
Inglese B2 ascolto · B2 parlato · B1 scritto
English B2 listening · B2 speaking · B1 writing
§07 — CONTACT

Costruiamo qualcosa
di affidabile, insieme.

Let's build something
reliable, together.

Email pinniola@hotmail.it
Sede Location Santeramo in Colle (BA) · IT
Disponibilità Availability Progetti selezionati Selected projects
Shortcut Premi ⌘K per i comandi Press ⌘K for commands