Hi 👋🏻, I'm Danilo Corsi.
🏢 AI Engineer & Technical Lead at NTT DATA Italia.
📜 Microsoft Certified: Azure AI Engineer Associate (AI-102).
👨🏻🎓 M.Sc. in Computer Science from Sapienza University of Rome (110/110 cum laude), specialized in Machine Learning, Deep Learning, Big Data, and Computer Vision.
🤖 I specialize in building and deploying production GenAI services, multi-agent architectures, and intelligent document processing pipelines. On the software engineering side, I build and integrate scalable microservices, hybrid search systems (dense vector + BM25), and resilient backends in cloud (Azure) and enterprise environments.
💼 Alongside core development, I act as Technical Lead: defining service architectures, running code reviews with automated testing, and aligning technical delivery with client needs.
🌐 Portfolio: corsidanilo.github.io/portfolio
-
✍️ Whisper Audio/Video Transcription App
Cross-platform offline transcription application using OpenAI’s Whisper. It allows transcription of audio and video files into text and enables analysis of the generated output using LLMs such as Gemini or locally run models via Ollama.
(Python, LLM, PyTorch, CUDA, Deep Learning) -
🤖 Multi-Agent Documentation Generator
Automatic, intelligent documentation generator using a multi-agent pipeline (7 specialized agents) and AST-based code analysis to produce standardized Markdown documentation, OpenAPI specs, and Mermaid diagrams.
(Python, Multi-Agent Systems, LLM, AST, CLI) -
📄 Simple Document Scanner
A high-performance, modern Android application built to simplify document scanning, processing, and sharing. Powered by Google Play Services ML Kit, this app provides a seamless professional scanning experience directly on your device.
(Kotlin, Android, ML Kit, Computer Vision) -
🎙️ Simple Transcriber
A privacy-focused Android application for high-quality audio transcription and text refinement, featuring both on-device AI for privacy and cloud-based AI for state-of-the-art accuracy. Built with Kotlin and Jetpack Compose.
(Kotlin, Android, ML Kit, offline, Huggingface) -
🗣️ Simple Speech App
A privacy-first, 100% offline real-time speech-to-text Android application powered natively by whisper.cpp with a system-wide floating accessibility overlay. Built with Kotlin, Jetpack Compose, and C++ NDK.
(Kotlin, Android, Jetpack Compose, C++, whisper.cpp, JNI) -
📌 Simple Shortcut App
A simple Android application that allows you to create custom home screen shortcuts. When tapped, each shortcut displays a scrollable grid of selected app icons for quick launching.
(Kotlin, Android, Jetpack Compose, Room Database, MVVM) -
🎥 Video Analyzer Utility
A Python utility to analyze videos by extracting frames using FFmpeg and querying Vision LLMs (both cloud like Gemini and local like Ollama or LM Studio) to get summaries, descriptions, or visual transcriptions, featuring an integrated AI assistant.
(Python, FFmpeg, Gradio, Gemini API, Ollama, LM Studio) -
📝 Text & Image Extractor Utility
A multi-format text extraction utility in Python to retrieve and analyze text from documents (PDF, DOCX, TXT, MD, CSV) and images using classical offline OCR (Tesseract) or advanced Vision LLMs.
(Python, Tesseract OCR, Gradio, PyMuPDF, Gemini API, Ollama)
-
🎛️ Leveraging LLMs for Informed Bitcoin Trading Decisions (Master's Thesis)
Exploring the use of Large Language Models (LLMs) to assist Bitcoin traders by analyzing sentiment from news and social media data.
(Python, Deep Learning, Pytorch, LLMs, Gradio) -
▶️ Multimodal Youtube Advertisement Sentiment Analysis
Implementation of deep learning models to perform multimodal advertisement sentiment analysis using tone of the voice, facial expressions, and remote photoplethysmography.
(Python, Deep Learning, Pytorch, Multimodal Interaction) -
🤼 Skin Lesion Classification
Implementation of deep learning models to perform skin lesion classification.
(Python, Deep Learning, Pytorch, CNN, Vision Transformers, StyleGAN) -
📦 SmartSupply
Implementation of a DApp using the Blockchain that allows customers to verify the authenticity of the products they purchase.
(Solidity, Typescript, Hardhat, Ethereum, Express, Docker, ReactJS, Prisma) -
🪙 Bitcoin Price Prediction
Implementation of machine learning models that try to predict the price of Bitcoin using features related to its price and blockchain information.
(Python, PySpark, Machine Learning) -
🍽️ Voicefork
Implementation of TheFork mobile application clone through a microservice structure.
(MobileApp, Typescript, Python, Docker, React Native, MySQL, PostgreSQL, Redis, Terraform, k6, Express, Prisma, FastAPI, Faiss) -
🗣️ Voicefork Alexa Skill
Voicefork add-on that implements an Alexa skill, allowing you to make restaurant reservations using your voice.
(Voice User Interface, Alexa Skill, Typescript) -
👋 BiometricBites
Web application that allows biometric access to university cafeteria with face recognition.
(WebApp, FaceRecognition, Typescript, ReactJS, SCSS, Django, Python, MySQL) -
🛩️ DroNET
Homework for the master's degree in Computer Science course "Autonomous Networking" at the University of Rome "La Sapienza," consisting of implementing a routing protocol for UAV ad-hoc wireless networks using Q-learning.
(Python, Reinforcement Learning) -
🫠 Stroke Prediction Model
Homework for the master's degree in Computer Science course "Fundamentals of Data Science" at the University of Rome "La Sapienza," consisting of creating a machine learning model for predicting strokes.
(Python, Machine Learning)
-
🎮 SanLo Defence
Development of a mobile tower defense game for the revaluation of the San Lorenzo district (Rome). Worked on the development of the frontend (main screen, main menu, and submenus).
(Frontend, C#, RestAPI, Unity, Figma) -
🚗 GeneroCity
Development of an Android app that allows users to easily find a parking spot. Worked on the frontend development (home screen, parking list, and some parking/unparking functionality).
(Mobile App, Frontend, Java, Android Studio, Adobe XD)