Carlo Morelli

Software Engineer

Milan, Lombardy, Italy21 yrs 2 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

20 years of engineering expertise in software and cloud technologies.
Co-authored 5 papers for IEEE journals and conferences.
Expert in ML/AI convergence within development lifecycles.

Stackforce AI infers this person is a SaaS expert with extensive experience in cloud-native solutions and quality engineering.

Contact

Skills

Core Skills

Machine LearningSoftware ArchitectureQuality EngineeringData LakesCloud ServicesDevopsJava DevelopmentTelecommunications EngineeringIntegrationResearchSignal Processing

Other Skills

Vector DatabasesContinuous DeliveryPythonMIMOMATLABVHDLReact.jsZigBeeLinux System AdministrationNI LabVIEWKubernetesMessage QueuePL/SQLArtificial Intelligence (AI)Agile methodologies

About

20 years of engineering expertise: software developer and architect, cloud native expert with AWS certifications, certified professional QA, and currently working ML/AI convergence in dev lifecycles; experience in multiple telecommunication topics, including networking, RF, digital signal processing and computer vision; co-author of 5 papers for IEEE journals and conferences.

Experience

21 yrs 2 mos

Total Experience

4 yrs 2 mos

Average Tenure

7 yrs 6 mos

Current Experience

Amazon

2 roles

Sr. Software Development Engineer

Promoted

Aug 2021 – Present · 4 yrs 9 mos · Milan, Lombardy, Italy · On-site

My main scope is to design, build and orchestrate tooling solutions and services that will help the Quality and Product management teams in form of account management services, virtual device and physical device farms (what we call “Lab as a Service”), and test automation frameworks that help automate the e2e Alexa+ and Alexa Classic testing for international scenarios. I am an Amazon Security Certifier since 2023 and Operational Excellence bar raiser in the business org advocating best practices concerning infrastructure, security and cost management. I also actively mentor junior engineers on coding and personal (functional and non-functional) development. Finally,
I am part of the European international charter of Alexa International Tech engineering Samurais.
What problems did I help solving?
Build and orchestrate at large scale a system that connects with multiple upstreams, performs ordering of ephemeral test accounts, books physical devices (from an internal Device farm), stores entitlement information and provides API for test execution frameworks, with an average offered rate of 5k requests per day;
Integrate the Lab-as-a-Service test framework with customers both on the API and on the Skills and engineer a plug-in system that scales; this has resulted in the last 2 years on > 10 plugins being added including contributions from external teams;
Study and use LLMs, RAG techniques and Vector databases (AWS Kendra or Elasticsearch) to cache RSS news feeds and provide evaluation frameworks for Alexa+ with up-to-date world news data;
Develop a Red-Teaming service for iterative guard testing of sensitive content of any target Large
Language Models, used internally to launch Amazon Nova models.

Machine learningVector DatabasesSoftware Architecture

Sr. Quality Assurance Engineer

Nov 2018 – Aug 2021 · 2 yrs 9 mos · Milan, Lombardy, Italy · On-site

Senior quality expert in the Alexa International Quality (AIQ) organization, the largest QA org inside Amazon, with the task of supporting Alexa expansion towards non-English speaking locales. I began my experience as Country Expert for the Italian market (and it-IT locale) -- with main duties are to orchestrate engineers for test sessions for the pool of Echo Family Devices / FireTVs and to build tooling for automation around the Natural Language Unit. After 2 years of extensive automation development, I was promoted as Tech Lead for the newly created Device & Architecture Team inside AIQ. The mission of this team is to build & architect solutions that allow to vendor AIQ
test artifacts to other Alexa teams, built to automate both NLU and Speechlet testing, being it audio-only or visual (for multimodal devices / 3p devices).
What problems did I help solving?
Create a sync platform with AWS serverless tools that allows to keep in sync JIRA defects and Alexa dev tasks without human intervention. This scaled to cover automatically about 10k updates per year, saving up to 10h per engineer per week
Build a data-lake solution for global reporting and metrics collection/processing inside AIQ organization, interconnecting various upstreams into an AWS Athena / SQL browsable set of tables; the data from the lake is then queried from Business Intelligence engineers in the org with Tableau dashboards that anyone can study
Built tooling that facilitated QA engineers’ life by scripting tedious activities and saved 5 to 10h / week for each team member.
Authored an automated solution to test NLU updates based on Release Notes standard format and
parsing, a solution that allowed to detect early defects in NLU releases in hours instead of a week of
manual test campaign.

Software ArchitectureData LakesQuality Engineering

Irdeto

Senior Software Engineer

Feb 2015 – Oct 2018 · 3 yrs 8 mos · Hoofddorp, North Holland, Netherlands · On-site

In Irdeto I was part of a fast-paced scrum team developing IrdetoControl, a next-gen Digital Rights Management (DRM) product with integrations on Google Widevine, Apple FairPlay and Microsoft PlayReady standards. IrdetoControl is a fully cloud-based solution to provide decryption keys to users of OTA/Streaming platforms using the mentioned standards, aiming at Millions of requests/second. Development makes extensive use of AWS cloud services and DevOps disciplines (infra-as-a-code, containers, provisioning of VMs, secure infrastructure, conf sharing, message queuing), Continuous Delivery pipelines and Telemetry and metrics.
What problems did I help solving?
Develop a watermarking solution for OTA/Linear TV broadcasting on MPEG2-TS standard with customers such as BeIN, Liberty Global
Develop a file solution offline watermarking for high-quality (50 Gb/movie) data for Warner Bros.
Creating and owning microservice for IrdetoControl to perform identity provider bridging to auth0 for
connecting users, aiming at a cumulative of a million of req/sec
Creating and owning microservices for key generation to each of the supported standard (e.g. PlayReady, Widevine and FP) of the IrdetoControl platform, and scripting the overall IrdetoControl solution in Kubernetes clusters; developed the whole CD pipeline for the solution and also developed a minikube variant for dev-testings on personal computers to optimise dev experience.

Cloud ServicesContinuous DeliveryDevOps

Ericsson

Technical Development Engineer

Dec 2012 – Feb 2015 · 2 yrs 2 mos · Milan, Lombardy, Italy · On-site

Full-time developer in Milano for the Java-based automation framework "SWAT", intended to be used by multiple product lines within Ericcson. The dev activity was split between framework infrastructure (instrumentation drivers based on GPIB or Ethernet protocols, and equipment-under-test APIs based on SNMP protocol) as well as pure test case scripting based on TestNG.
In parallel, I was directly involved in deployment of performance testing framework "FAST" based on LabVIEW 2012 with G# toolkit (OOP extension).

Java DevelopmentTelecommunications Engineering

Nokia siemens networks

Integration & Verification Engineer

Nov 2005 – Oct 2012 · 6 yrs 11 mos · Milan, Lombardy, Italy · On-site

I was a specialized integrator of packet networking for mobile backhauling, interworking the NSN FlexiPacket packet radio product with NSN, Cisco and Tellabs routers and switches for last-mile transport, including synchronization and pseudo-circuit emulation support, and integration
with many OEM products: FlexiHybrid, PowerHopper, Ceragon IP10, Dragonwave Horizon+.
As R&D engineer I often travelled for product support on other company sites (Espoo, Shanghai, Tel Aviv, Munich) and for acceptance testing at the customer premises (New Delhi, Newbury).
I contributed designing the platform FlexiPacket Lite for micro-cellular LTE backhauling, based on PHY layer of IEEE 802.11n (OFDM MIMO) and proprietary MAC layer logic. I was in the automation team that developed the RF performance-testing platform based on LabVIEW 2009 within the company. Part of the automation stack was also based on Python 2.x and TCL/TK.

PythonMIMOTelecommunications EngineeringIntegration

Politecnico di milano

Contract researcher

Nov 2004 – Oct 2005 · 11 mos · Milan, Lombardy, Italy · On-site

My activity was centered on research about localization algorithms based on Bayesian techniques (Hidden Markov Models, Kalman Filters, Particle Filters and Jump Markov Filters) over diverse infrastructures like Wi-Fi, UWB, ZigBee or MOTEs.
I was co-author of 3 conference papers and 2 columns published on IEEE Transactions on Signal Processing magazine, co-advisor for two BSc and one MSc master theses, and system operator of a medium-sized thesis lab with 8 Linux/Windows machines plus 1 Solaris Ultra40.
Technologies: Matlab 7, C++, VHDL, Linux, Solaris
Platforms: RFID, UWB, ZigBee, Motes

MATLABVHDLResearchSignal Processing