In the team of Amazon Q: the generative AI-powered assistant

Working with a multidisciplinary team on language detection; Human Hall is backed by European Union and MUSA, with several partners like Deloitte

- Co-directed the development of pipelines for fine-tune LLMs and synthetic data.

Tech: Python, Hugging Face, Ollama, Unsloth, AWS

Robotics. Research team mainly composed of PhD Applied Scientists. Mission: develop robots to pick any kind of item from human-task-designed pods in warehouses. IEEE

- Designed from scratch a flexible, modular and extensible architecture for the operators UI, to speed up the development pipeline, and to enable external testing.

- Took ownership of the project and worked side by side with stakeholders; suggested and guided the refactoring the whole team data schema

- Winner the Amazon Robotics internal hackathon with a project about streaming and remote controlling the Universal Robots

Tech: React, TypeScript, ProtoBuf, Kotlin, Python, AWS

FamaLabs is an experienced company in distributed systems and cloud architectures in healthcare.

- Took ownership and proactively guided the development of the modular survey engine and compiler, and designed UI library for MAP (Medical Adaptive Platform): a dashboard built for doctors to easily manage CRF and more.

Tech: Next.js, TypeScript, Redux, Jest, Storybook

Founding team. Worked on War of Wheels: a multiplayer game about medieval-punk fighting cars.

- Winner of BGF: funding of 120k and starup accelerator; participated in several international events (like GamesCom Cologne, GamesConnection Paris and GDC San Francisco) while pitching to publishers and investors.

- Helped the growth of the team from 5 to 20 employees; proactively leaded the team for the design and implementation of the multiplayer and networking system.

Tech: Unity, CSharp, PhotonEngine, Python

Smart-Bear is a big data platform in healthcare funded by the European Union and in collaborations with top European Universities; it offer evidence-based interventions thanks to predictive analytics.

- Built modular and cloud-optimized pipelines for predictive analysis, data imputation and visualization while protecting user health data anonymity.

Tech: Python, PySpark, Echats.js and AI/ML libraries

Developed several commercial projects from videogames to AI automations and websites. Here are some of the most relevant:

- Untitled Nature Game: a prototype game that went viral with several millions of views across TikTok, Reddit and Twitter. It got offers from several investors and publishers.

- RELOW: a solo-developed multiplayer game, released in 2020 with 5k+ sales, showcased in several expositions like Milan Games Week 2019 and 2020, and Svilupparty 2019 and 2020; attended and given several talks and interviews.

- POLYWAR: multiplayer game released in 2016 with 200k+ sales in collaboration with the publisher New Reality Games

- Easy Insta Growth: 2020-2021 an Instagram AI automation, with hundreds of clients worldwide.

- Freelance worked on several websites

Tech: Unity, CSharp, Python, React