The startup is focused on A.I. powered products, I'm a member of the team of 4 data scientists and 3 developers, working on the first product of the company: a Recommender System for E-Commerce websites. The team activities are managed with Scrum and there is a shared DevOps view of our work.My responsibilities:- creation of the development environment for the team (Docker, Docker-Compose, Traefik)- automatic building of Machine-learning Models using Airflow (Google Cloud Composer) - deployment of Machine-learning Models as microservices using KubeFlow/Seldon-Core- automation of image-building and testing on git updates (GitLab CI, Slack)- deployment on Kubernetes (GKE on Google Cloud)- automation of monitoring and metrics gathering of the application API (Prometheus, Express-Gateway)

My core activities are all around the company platform Volare, an integrated solution to offer guest-wifi and Location Analytics, used every day from thousands of people in all the globe.Responsibilities:- ordinary activities on infrastructure used for Public Clouds offering (3 AWS regions) and in private datacenter for customers that opted for on-premise setup;- delivery of new Volare instances for Enterprise customers: architectural project, meetings with customers to match internal requirements, security checklist.- support dev team improving the architecture of the platform and the capacity to be monitored, scaled and operated.Beside of the daily activities, there are projects that I cultivate with mid-long term commitment:- adoption of Docker as a company standard to deliver all the components of the platform. This required an initial training to the team to reduce to technical debt, then reorganization of the source code, centralization of configurations using Consul by Hashicorp and logging on AWS CloudWatch; supported team to realize production-ready Dockerfiles for the different programming languages in use (PHP, Java, Golang, Python, NodeJS).- introduction of a development flow with shared view of activities through the creation of a CI/CD Pipeline: automatic build of Volare components images on branch and on tagged releases, automation of tests during the build of the APIs and automatic deploy in test environment. Tools used: Docker, Gitlab, Slack, Postman/Newman, Docker-Compose.- evolution of the monitoring system, from the periodic polling approach on each host to the push telemetry system, capable of recognizing not only timely failures but measuring the consumption of internal resources for the individual components. Tools used: Logstash, StatsD, Graphite, Grafana, CollectD.- support the Machine-Learning team in features extraction from raw data and production-ready workflow management (FileBeat, LogStash, Airflow, Slack)

I've been the Head of System Administrators department, reporting to the CEO and coordinating the job of the ops team (up to 7 people).Responsibilities:- coordinate ordinary activities on physical datacenter infrastructure (5 racks in Telecom Italia Bologna, 3 racks in BT/I.Net Milano), responding to SLA requirements of Enterprise customers and internal Quality Assurance Office;- project and manage IT infrastructure development, mainly focused on hosting web applications with high traffic (newspapers, e-commerce, sports & fashion events);- coordinating support for company offices (Prato, Milano, Cascina (PI))Main projects:- take in charge of about 20 racks of physical servers from the controlling group Dada (2007);- massive virtualization of application cluster, reducing from 20 to 8 racks the space required to offer the same services (2008-2009, using Xen/XenServer);- physical migration of systems from Dada dedicated spaces to Telecom Italia datacenter in Bologna, without service interruption of web sites hosted (2007-2009);- migration from internal solution for groupware (Zimbra + Exchange) to Google Apps for Work;- introduction of a trouble ticketing system to trace internal & customers requests and project management based on Wiki and SVN (2007);- guided transition from manually managed infrastructure to "Infrastructure as Code" approach using Puppet (2009);- creation of a clustered hosting solution based on web control panel IspConfig, allowing self-provisioning of the tecnical teams; the system was completed with a charge-back system to reconcile hosting costs to the revenue activities (2008-2010);- Inbox zero project: creation of company intranet in collaboration with marketing and QA offices, for information sharing and the reduction of email overload (2012);- setup of a Privacy management system regarding all the projects delivered by the company aiming the delivery of services "compliant by design" (2014);

Software developer on LAMP Platform for medium/big-sized sites, mostly belonging to Dada or RCS group

I've worked on small & medium size web projects, with Non-Profit & Government customers.