Davide J. Viero

Backend development and cloud-native infrastructure engineering, including LLM operations.

• Backend development (TypeScript, AdonisJS)

• Google Cloud Platform DevOps (Kubernetes, Pulumi, Vertex AI)

• LLM DevOps (RunPod, vLLM)

• Architecture, sizing, development, usage models

Development of internal tools for the management of the company's workflow.

• NodeJS/PHP backend (Fastify, AdonisJS)

• Vue/MD Lite frontend

• GNU/Linux SysAdmin, Docker

Part of the dev team of different Consonant projects (both for internal use and for external clients) as a FullStack developer.

• Node.js backend (FeathersJS, Express)

• Vue + Vuetify, Vue + Quasar frontend

• PostgreSQL, MongoDB, ElasticSearch/OpenSearch databases

• Wireguard

• OpenCV + depthAI on OAK-D cameras for Computer Vision tasks

• AI toolkits for LLM and RAG: LlamaCPP, langChain, LangFlow, Vector Databases

• System Networks Admin: IPTables, NFTables, custom router build

Analysis of the capabilities and performance of Apache Spark in conjunction with the AXS library about cross-matches between big astronomical catalogues in the Euclid Archive System (EAS).

• Installation and configuration of Apache Spark in a EAC (European Astronomy Center) cluster

• Crossmatch between different astronomical catalogs (Gaia DDR2, ALLWISE, KiDS, Euclid flagship)

• Comparison of the obtained data with the crossmatches run on different platforms (Greenplum, Gaia Archive)

• Results report & data correctness analysis

Development of a distributed fault-tolerant framework for digital signage using extensively Docker and RabbitMQ. Configuration of the systems (OS, network, storage) where the framework needed to be executed. Development of the backend infrastructure, using Docker, RabbitMQ, and Redis.

• Distributed fault-tolerant framework for digital signage

• Docker, RabbitMQ, Redis

• Systems configuration (OS, network, storage)

Freelance developer during the Erasmus experience. Most notable projects:

• Python video stream recorder

• C++11 raytracer

• Simple particle system in OpenGL

Development of the KU desktop application, an Electron based porting of the KU player from RaspberryPi to Windows and MacOs. Polymer was used for the UI and Three.js for the WebGL mapping.

• Electron desktop application

• Polymer UI

• Three.js WebGL mapping

Development of the KU player, an easy-to-use mapping player built with the RaspberryPi platform in a multi-device environment.

• Python for network software

• Node.js for main control software

• C with Arduino library for firmware

Development and maintenance of a TCP scheduler application for the "Tree of Life" EXPO 2015 installation in Milan, using Polymer and FullCalendar. General sysadmin work.

• TCP scheduler application

• Polymer and FullCalendar

• General sysadmin work

1° ciclo - Laurea L [not completed]