Speaker

Stefano Tempesta

Stefano Tempesta

Web3 Architect | AI & Blockchain for Good Ambassador

Gold Coast, Australia

Stefano Tempesta is a technologist working at the crossroad of Web2 and Web3, to make the Internet a more accessible, meaningful, and inclusive space. Stefano is an ambassador of the use of AI and blockchain technology for good purposes. A former advisor to the Department of Industry and Science, Australia, on the National Blockchain Roadmap, he is co-founder of Aetlas, a climate fintech integrating blockchain-powered forward-buying contracts into fully-digitised green bonds, simplifying and automating project financing for Carbon Removal infrastructure.
Stefano is a lecturer at RMIT University on courses about AI and blockchain technology, and he has co-authored the book "Blockchain Applied".

Area of Expertise

  • Information & Communications Technology

Topics

  • Enterprise Blockchain
  • Web3.0
  • Machine Learning and Artificial Intelligence

Protect your data with SQL Server Always Encrypted with secure enclaves

Always Encrypted with secure enclaves expands confidential computing capabilities of Always Encrypted by enabling in-place encryption and richer confidential queries.
This session shows how Always Encrypted protects the confidentiality of sensitive data from malware and high-privileged unauthorized users: Database Administrators (DBAs), computer admins, cloud admins, or anyone else who has legitimate access to server instances, hardware, etc., but shouldn't have access to some or all of the actual data.

Introducing London Bridge, the secure bridge between Algorand and Ethereum

Using Intel SGX hardware secure enclave technology to verify each blockchain, London Bridge is a unique technology to send transactions and manage keys between the Algorand and Ethereum networks.
This session describes the technology at the foundation of London Bridge, and shows an example of how to transfer tokens between Algorand and Ethereum using the additional security of State Proofs and Hardware Enclaves.

Discover the Azure OpenAI Service responsibly

Azure OpenAI is a new AI service in Azure that provides a REST API to access the OpenAI's powerful language models including the popular Chat-GPT.
These APIs can be easily adapted to specific tasks, including but not limited to content generation, summarisation, semantic search, and natural language to code translation.
This session will provide a panoramic of such services and how easily is to program Azure OpenAI in a variety of programming languages.
In addition, this session explores more in detail the content filtering and moderation service. Azure OpenAI includes a content filtering system that detects and takes action on specific categories of potentially harmful content. The examples provided in this session will have an emphasys on ensuring that coding and language AI models are used responsibly, for their intended purpose.

Carbon Asset Solutions plants the seed for net zero emissions securing carbon credits with Azure

Carbon Asset Solutions enabled Azure confidential computing in its IoT devices and data management for recording and verification of carbon sequestered in soil. Azure confidential ledger allows CAS to efficiently verify and securely sell carbon credits to the rapidly growing net zero carbon markets.
This session presents technologies for an effective carbon sequestration program: precisely measure carbon in soils, an immutable record of data, and an ordered verification system to track credits.

Walking Between Worlds, connecting indigenous artists to blockchain and NFTs

Deep dive into the technology that runs "Walking Between Worlds", how indigenous artists from the Aboriginal communities of Australia, a 50,000+ years old culture, are embracing blockchain, Web3 and NFTs for selling artwork and raising funds for charity.
Topics covered: artwork rendering, rarity calculation, metadata generation, publishing to IPFS, integration with OpenSea, ERC-721 smart contract in Ethereum, web3.js and minting process, Web API and smart contract interaction with NEthereum.

ML for Anti-Money Laundering

Effective Anti-Money Laundering (AML) solutions can help financial services fight global crime, including human trafficking. Traditional approaches to AML consisted of event scanning based on risk tables. The adoption of ML techniques has helped organizations scale AML to a broad range of data sources and transaction data in real time. Still, ML-only approaches bring inevitable bias that may hinder minorities.
This session explores the technical design and the ethical considerations for an AML solution to extract and ingest data from watch lists and multiple data sources, making data available quickly for analysis, and improve accuracy in detecting AML activity while reducing wasted effort in investigating false alerts. Design and implementation of the ML components of the anti-money laundering solution are also seen in respect of the principles of responsible AI.

Explore the Microsoft Graph for Teams

Microsoft Graph is the gateway to data and intelligence that provides a unified programmability model that you can use to access the tremendous amount of data in the Microsoft 365 platform.
This session presents the specific Graph endpoints for Microsoft Teams, and shows working examples of how to create a new team, add people to the team, and configure a team with channels, etc. All done programmatically, in your preferred programming language!

Serverless AI with Custom Vision & Azure Functions

This session describes the architecture of the Custom Vision service to build ML image classifiers, to bulk label images with a REST API, and to train a model and use it online and offline in a mobile app. Custom Vision provides developers with the possibility to personalize the image detection and analysis service of Azure Cognitive Service, to tailor the data model to specific use cases. Trained model can be utilised also in environments with no live connectivity to the cloud, offering the possibility to deliver intelligence at the edge. The session describes also the continuous learning of the ML classifier based on newly collected images, and how to create a time series of images persisted in Azure storage using serverless functions.

AI-powered SharePoint Intranets

This session combines the agility of building pages and web parts in SharePoint Framework, with the power of the Microsoft AI platform. Specifically, I'll present a dashboard in SharePoint that displays Machine Learning-powered sentiment analysis of your intranet contributions; an automatic document and image classification with Cognitive Service, and a content filtering engine that learns from new entries and improves accuracy of detection over time.

Content Moderation in Microsoft Teams

Content Moderator is a machine-assisted content moderation service and human review tool for images, text and videos. With Content Moderator you can detect potential offensive images, filter possible profanity and moderate adult and racy content in videos. Content Moderator also checks for personally identifiable information (PII).
This session shows how to moderate content uploaded in Microsoft Teams, filter out unwanted images through machine-learning based classifiers, custom lists and optical character recognition, and how to detect undesirable text in a variety of languages.

Protect your financial ML.NET workloads with Confidential Computing

Preserving privacy when processing data from multiple sources with machine learning is always a challenge. Organizations may want to perform collaborative data analytics while guaranteeing the privacy of their individual datasets. Combining multiple data sources to support a better algorithmic outcome improves accuracy of prediction, but it may come at cost of confidentiality, if sensitive information is not accurately protected.
Azure Confidential Computing adds new data security capabilities to the Cloud and specifically to machine learning processing. By using trusted execution environments (TEE) to protect your data while in use, with confidential computing, you can use machine learning algorithms across different organizations to better train models, without revealing the processed data.
This session presents the benefits of confidential computing in a Machine Learning solution, where different financial institutes share their confidential datasets for data analysis and credit risk prediction using the ML.NET library, and still mask any sensitive information to protect the privacy of their customers, preventing any potential data leakage.

Minecraft AI for Good: Prevent Forest Fires

Yes, you read it well. This session is about Minecraft and application of AI technology! A village needs your help to prevent the spread of a nearby forest fire. By starting with an overview of Minecraft Education Edition, you will learn how to develop agents that prevent forest fires: Train agents to identify what causes fires, remove materials that help fires spread, and then bring life back to a forest destroyed by fire.

Virtual Eye Vision with Hololens

Virtual Eye is a vision technology that provides virtual sight to the visually impaired by recognizing people and the environment around using AI and AR capabilities of Azure. This session explores the technologies implemented along the HoloLens device to perform body tracking and image recognition in real time, and translate this information in vocal instructions that describe the world around the user.

Measure your Teams sentiment

Sentiment analysis is a special case of text mining that is increasingly important in business intelligence and social media analysis. In this session, we’ll build an experiment for sentiment analysis of conversations in Microsoft Teams to measure the overall feeling of engagement of your teams. For example, sentiment analysis of comments can help organisations monitor appreciation and utilisation of their IP (Intellectual Property), or help users identify opinion polarity before accessing a resource. The output predictions can be aggregated over specific tags containing a certain keyword, in order to find out the overall sentiment for each element of the taxonomy, and lastly consumed by analytical tools such as Power BI.

An Architecture Reference for Confidential Decentralized Identity

The identity & credentials economy has been relatively stable for a very long period. .For hundreds of years, it has been a quasi institutional monopoly centered around governments and universities, and with relatively little change and innovation. But recently, this industry has begun to be disrupted by the globalization of higher education and labor markets, requiring credentials to work over a much larger scale. The increasing need for access to digital services, which require multiple forms of digital identity, for digital wallets, online banking, social media accounts, etc. carries significant risk, if not thoughtfully designed and carefully implemented.

A new form of identity is needed, one that weaves together technologies and standards to deliver key identity attributes, such as self-ownership and censorship resistance, that are difficult to achieve with existing systems. Cryptographically secure, decentralized identity systems could provide greater privacy protection for users, while also allowing for portability and verifiability.

This session describes an architecture reference for Decentralized Identities (DIDs) that fits in the self-sovereign identity (SSI) framework of issuer - holder - verifier process. The solution architecture demonstrates how to make this "trust triangle" trustworthy and confidential. Confidential Identity Hubs, hosted on confidential computing infrastructure in multiple “hubs” around the world create the necessary distributed network for storing and securing identity elements at scale. Security of sensitive data is ensured by the redundancy of the distributed network, and governance remains decentralized by removing sole ownership of the provided infrastructure.

Biometric Security in ASP.NET Core with Cognitive Services

The new ASP.NET MVC Core framework introduces a claim authorization mechanism that accepts custom policies to restrict access to parts of your web application depending on the current authenticated user.
This session introduces the new policy-based model to decouple your authorization logic from the underlying user roles, and presents a specific usage of such authorization policies based on biometric information, such as face or voice recognition.

From Spaghetti to Microservices Architecture

Hopefully, far has gone the time when systems were built like monolith and integrated with point-to-point connection… right? More likely, though, still many software applications are developed with a convoluted design that, eventually, will hit the wall of maintainability and scalability. In this context, how can a microservice-based architecture help organisations focus on building features that add business value to their applications, without the overhead of designing and writing additional code to deal with issues of reliability, scalability, or latency in the underlying infrastructure?
This session explores the agility of architecting fine-grained microservice applications that benefit of continuous integration and development practices, and accelerated delivery of new functions into production, with the help of Azure Service Fabric. It also presents the Publish-Subscribe design pattern of an enterprise-level service bus built on Azure Service Bus, which guarantees message queueing and delivery, on-premises and in the Cloud.
Targeted at software architects and developers, during this session, a significant emphasis is posed on demoing the ESB capability available in Azure, how to avoid spaghetti-like intricate architecture designs, and how to design for microservices and API-based applications.

Exploring CQRS and Event Sourcing: A journey into high scalability and availability with Azure

The Command Query Responsibility Segregation (CQRS) pattern and Event Sourcing (ES) are currently generating a great deal of interest from developers and architects who are designing and building large-scale, distributed systems.
This session is structured in two related parts: An introduction to CQRS and ES to get you started with the pattern and domain of application, requirements, potential barriers; and a working reference implementation sample built in Azure with Service Fabric, which is intended to illustrate many of the concepts related to the CQRS pattern and event sourcing approaches to developing complex enterprise applications.

Improve your email sentiment using Machine Learning in an Outlook add-in

Emails represent an important means of communication, formal and informal, within and among organisations. The lack of visual and emotional feel, however, make emails prone to misunderstanding and inappropriate use of tone of expression. Can we use modern technology to improve our communication? This session presents an “experiment” of analysing email sentiments and identifying patterns of expressions that influence such sentiments, with the purpose of providing an immediate feedback to the user before sending the message. The adoption of this feedback add-in in Outlook has improved the sentiment of communication within an organisation, addressing cases of untimely emails, disrespectful messages and biased communication.
During this session, we will build an Outlook add-in that uses Microsoft Cognitive Services Text Analytics API to analyse the sentiment of your email contents, and give you feedback. Feedback is then collected in a database, anonymously, and patterns are identified based on various criteria, including department, time periods, gender, reporting line, response time, etc. Information is then displayed in a Power BI dashboard.

Microservices in practice: Design, platforms, tools and code

Microservice architecture decomposes monolithic applications into discrete, atomic, full-stack service silos. It’s patterns and principles take over where Service-Oriented Architecture left off, by including recent advances in containerization and DevOps culture. In fact, embracing Microservice architecture forces you into better practices related to design and operations, including Domain-Driven Design, single responsibility pattern, Command Query Responsibility Segregation (CQRS) and Event Sourcing (ES).
This workshop begins with a journey through the evolution of Microservices principles and patterns, followed by design discussions of real applications and visual topologies, and by a deep dive into platforms and tools. You’ll learn the process of design, development, deployment, upgrade, scale and operations for Microservices platforms and Container-based solutions on the Microsoft Azure Cloud.

A pizza ordering bot in 30 minutes, live on stage. Pizza not included!

Whether you believe it’s the future or not, artificial intelligence is a hot topic today. Microsoft announced its long-term vision on AI, and Cognitive Services and the Bot Framework are part of this strategy.
In this session, I’ll build a pizza ordering bot with the Microsoft Bot Framework in 30 minutes, live on stage. No pressure, challenge accepted! We’ll see what makes a conversational bot a “great bot”, how to connect to channels for an improved conversation experience, meet existing bots and enable them to interpret and interact in a human way. And we’ll enhance the user experience with actions like greeting customers after they start chatting, or providing a personalised menu. Our aim is to get the bot deliver our favourite pizza by thinking that we are interacting with an actual human being. Pizza not included, sorry!

Modern Software Security Development Lifecycle

Modern software development processes require software engineers to design and build more secure software and address security compliance requirements while decreasing development cost. Reducing the opportunities for attackers to exploit a potential weak spot or vulnerability requires analysing the overall attack surface, and includes restricting access to system services. Applying a structured approach to threat scenarios during design helps a team more effectively and less expensively identify security vulnerabilities, determine risks from those threats, and establish appropriate mitigations.
This session illustrates the core concepts of the Microsoft Security Development Lifecycle (SDL) and discusses the security activities that should be performed in order to claim compliance with the SDL process. Combining a holistic and practical approach, the SDL aims to reduce the number and severity of vulnerabilities in software by introducing security and privacy throughout all phases of the development process.
Besides presenting the Microsoft SDL methodology, this session presents practical applications of tools for understanding your attack surface before and after new apps are deployed (Attack Surface Analyzer), finding and addressing system security issues (Microsoft Threat Modeling Tool), and a simple fuzzer designed to test for potential denial of service vulnerabilities (MiniFuzz).

Scaling applications with Azure Redis Cache and Machine Learning

In a multi-tier application, bottlenecks may occur at any of the connection points between two tiers: business logic and data access layers, client and service layers, presentation and storage layers, etc. Large-scale applications benefit of various levels of caching of information for improving performance and increasing scalability. Caching can be configured in memory or on some more permanent form of storage, in different size and in diverse geographic locations. The open source Redis engine, as implemented in Azure, allows for an intuitive configuration of management of all these aspects, and utilisation from a variety of programming languages.
At EF Education, our applications are used by hundreds of thousands of students and staff members daily in 150+ locations world-wide. How do we scale to this mass? How do we optimise performance across regions? This session presents design best practices and code examples for implementing the Azure Redis Cache and tuning the performance of ASP.NET MVC applications, optimising cache hit ratio and reducing “miss rate” with smart algorithms processed by Machine Learning, and for automating and monitoring the deployment of the Redis cache across different tiers, persistence layers and replicated nodes.

Effective emergency response with Azure IoT and Cognitive Services

Schools at EF Education First handle thousands of students every week in more than 150 locations worldwide. Emergencies may happen with no notice, whether for weather-related events or a terrorist attack. How do we react promptly and safeguard the security and safety of our students and staff around the world?
A system of multiple communication channels is used to reach out on students and inquiry about their incolumity. The last known location is tracked with GPS units and its data collected and analysed via the Azure IoT Hub; automatic messages and calls in multiple languages are initiated from the CRM and replies processed with Azure Cognitive Services using text and voice recognition and translation; bots are used for checking status and condition.
Targeted at software architects, developers and product owners, this session explores the core capabilities of the Azure IoT and Cognitive Services in providing an integrated and effective solution for immediate response to emergencies using a variety of communication channels and languages.

Ranking CRM leads and interests with Azure Machine Learning

Every day, EF Education First receives thousands of expressions of interest by prospective students to attend an education program delivered in any of the 150+ locations around the world. How is all this information processed promptly in order to provide a swift and effective response to applicants? We rank leads and interests based on program, location , past history, and hundreds, literally, of other criteria. We cannot do this manually clearly. We use the power of outcome prediction algorithms in Azure Machine Learning.
Targeted at software architects, developers and product owners, this session explores the foundation of Azure Machine Learning for building outcome prediction services, describing how data is collected and defined into a model; how the model is trained and then scored; and finally how the evaluation of the model is processed to generate the ranked outcome.
Custom decision-tree algorithms are presented in the programming language R, along with RESTful Web Services consumed by our CRM application. This session completes also with the illustration of best practices and guidelines for maintaining and deploying large-scale datasets in the Cloud and optimisation of computing time of ML experiments.

Bring Intelligence to the Edge with Custom Vision

This session describes the architecture of the Custom Vision service to build ML image classifiers, to bulk label images with a REST API, and to train a model and use it online and offline in a mobile app. Custom Vision provides developers with the possibility to personalize the image detection and analysis service of Azure Cognitive Service, to tailor the data model to specific use cases. Trained model can be utilised also in environments with no live connectivity to the cloud, offering the possibility to deliver intelligence at the edge). The session describes also the continuous learning of the ML classifier based on newly collected images.

Solve the Rubik’s cube with Azure AI technologies

Can you solve the Rubik's cube? Need help? Ask Kubik!
Powering Microsoft Cognitive Toolkit, an easy-to-use, open-source toolkit that trains deep learning algorithms to learn like the human brain, Kubik is a free API that solves the Rubik's cube by analyzing thousands of combinations in real time. Exposed as a public REST API, the service improves the resolution path of the cube's puzzle by constantly learning new and more optimized moves.
Additionally, the Kubic framework implements image recognition from Microsoft Cognitive Services for interactively observing a physical Rubik's cube and visualizing suggested moves to resolve the puzzle.

DevOps for Blockchain Smart Contracts

Blockchain has emerged from the shadow of its cryptocurrency origins to be seen as a transformative data technology that can power the next generation of software for multi-party enterprise and consumer scenarios. With the introduction of blockchain technology in enterprise software development, organizations are asking for guidance on how to deliver DevOps for blockchain projects. Blockchain applications are often designed to handle financial transactions, track mission-critical business processes, and maintain the confidentiality of their consortium members and the customers they serve. As a result, blockchain applications usually demand a more rigorous risk management and testing strategies than traditional software applications.
This session looks at core aspects of DevOps with a focus on best practices and tools to incorporate continuous delivery, continuous improvement, and infrastructure as code to the development of smart contracts for blockchain solutions in Azure.

Azure MLOps: Bring your DevOps to Machine Learning

MLOps empowers data scientists and app developers to bring ML models to production. This session presents how to use MLOps in Azure to track, version, audit, certify and re-use every asset in your ML lifecycle and streamline the use of each resource. With practical examples on asset management and orchestration services for your ML model training and deployment workflows, you will learn about the Azure DevOps Machine Learning extension, best practices for data scientists to work in topic branches off master, when code is pushed to a Git repo, how to trigger a CI (continuous integration) pipeline, and how to provision ML workspaces, compute targets, datastores as infrastructure-as-code.

Run your Supply Chain in the Azure IoT and Blockchain Cloud

Supply chain processes often require information collected in the field to validate transactions. Think of measurements of conditions of shipped goods, including storage temperature or humidity. This data is collected by different sensors at each stage of the supply chain, and validated in an immutable registry, represented by a blockchain digital ledger.
In this session, Stefano Tempesta, Microsoft Regional Director and MVP on AI and Business Applications, will demonstrate how to easily build supply chain workflows in Dynamics 365 that integrate IoT devices with a blockchain network using Logic Apps and the Ethereum Blockchain Connector. Relevant entities and data are stored in the Common Data Service (CDS). Transactions are validated by the execution of the associated smart contract on the blockchain that represents the appropriate projection of the entity that was added to the Common Data Model.

Secure multiparty Machine Learning with Azure Confidential Computing

Preserving privacy when processing data from multiple sources with machine learning is always a challenge. Organizations may want to perform collaborative data analytics while guaranteeing the privacy of their individual datasets. Combining multiple data sources to support a better algorithmic outcome improves accuracy of prediction, but it may come at cost of confidentiality, if sensitive information is not accurately protected.
Azure Confidential Computing adds new data security capabilities to the cloud and specifically to machine learning processing. By using trusted execution environments (TEEs) to protect your data while in use, with confidential computing, you can use machine learning algorithms across different organizations to better train models, without revealing the processed data.
In this session, Stefano Tempesta, Microsoft Regional Director and MVP on AI and Business Application, will present the benefits of Azure Confidential Computing in an ML scenario, where two separate health institutes collaborate on data analysis and prediction using Azure Machine Learning, and still mask any sensitive information to protect the privacy of their patients.

Electronic Signature with Smart Contracts and the Azure Blockchain Development Kit

The release of the Azure Blockchain Development Kit represents a milestone in the adoption of blockchain technologies in the enterprise space. Thanks to the Blockchain Development Kit, you can now build solutions that seamlessly integrate blockchain with the best of Microsoft and third-party software applications. Blockchain Development Kit works in combination with Azure Logic Apps and Flow to dramatically simplify the development of end-to-end blockchain applications that access on- and off-chain data, handle events generated by the digital ledger, and leverage the Azure ecosystem for a seamless and integrated solution.
In this session, Stefano Tempesta, Microsoft Regional Director and MVP on AI and Business Applications, will describe how to automate document sign and verify workflows in SharePoint using Azure Logic App and Azure Blockchain Workbench for persisting files’ hash and metadata on a blockchain digital ledger.

Azure Day Roma 2023 Sessionize Event

June 2023 Rome, Italy

Virtual Azure Community Day Sessionize Event

July 2020

Power Platform Fest : Summer 2020 Blast! Sessionize Event

June 2020

DeveloperWeek Global 2020 Sessionize Event

June 2020

Azure Day Rome 2020 Sessionize Event

June 2020

MVPDays Online January 2019 Sessionize Event

January 2019 Calgary, Canada

Global AI Bootcamp Sessionize Event

December 2018 Sydney, Australia

Stefano Tempesta

Web3 Architect | AI & Blockchain for Good Ambassador

Gold Coast, Australia