(MegaBlog) Microsoft Build: AI Day 2024 | EP: 1— A lot of Things Happening
By Charunthon Limseelo and Collaborators
Finding my photo album, please visit here >>> https://photos.app.goo.gl/oYkuXytzhwhq4SaB9
Hi developers, welcome to our blog again…
You may have seen the previous episode (Episode 0) that I came very early for one day before the event started, that’s just the rehearsal for me. But today is different, this is the actual event day…and I have some kind of energy coming, electrifying, entertaining, on the way from Seattle to Bangkok. I’m very thrilled much since it’s my very first international event for me to attend in. But first, I gonna introduce you what this event is all about.
Microsoft Build: AI Day Tour is a free, one-day event that helps developers (even students, software engineers, data scientists, architecture, and much more) discover new opportunities with AI and advance their knowledge and skills to deliver more value faster with AI and Microsoft Azure. Also, you can join local experts and peers to share ideas and unleash creativity with the power of AI.
During this event, you’ll be able to:
- Discover new opportunities with generative AI and be the first to learn about the latest innovations.
- Connect with Microsoft experts.
- Explore how to infuse responsible AI practices throughout your development lifecycle with state-of-the-art tools.
In this blog, I’m going to sum all up on what happened in this event and telling some new innovations which is shown from the event.
First, I can’t sleep…and I got up too early
On the night of April 30th, I could possibly say that I couldn’t have much good sleep because I kinda being excited from “Satya’s Energy”. So that, I got up at 4 am. and it’s crazy to wake up on that time. After that, I started travelling to QSNCC by skytrain at haft past 5, and landed at Asoke Station at 5:45. Found out that MRT is opening at 6 and I come toooo fast for this event. After MRT opened, I sat on MRT and then going to QSNCC. However, I can’t enter the area since the security guards said that the conference area (including shops and restaurants) is opening at 7, and I arrived at 6:15. You may gonna ask me where I would go before the hall opened. Bechakitti Park? The Parq?
Well, it’s none of them. Luckily, I have my Microsoft Learn Student Ambassadors Badge keeping with me and then accidentally entering to the convention center by the permission from security guards. Then, I went to find something to eat for my breakfast. So, I decide to go to Casa Lapin for having Lemon Ice Tea and a piece of Bagel. At that moment, I met P’Kom Supasate and N’Arm Siwakorn (both of them are CPE KMUTT like me) as the first group coming in the event. After I got my drink and snack, we went up to Plenary Hall for finding a place to take a rest.
Morning Time Until Late Morning…
At 7 a.m. — Time to setup the event booth
As three of us sat and took a rest for a while, I decided to walk around the lounge and took a bit sightseeing around the event. In meantime, first two Microsoft seniors just arrived to the lounge and have a conversation with. The first person who I would like to introduce to you and would be appeared at the session “What’s New in Generative AI” which joining with P’Si Mullika, his name is P’Jimmy Danai. And another person is the one who I really know her well, more than her brother, her name is P’Ploy Natcha, our former Micorosoft Student Partner.
One of the funniest moment before we were entering the event is how we saw P’Jimmy enjoyed riding (he said he’s walking) with his AirWheel around the lounge. And in the afternoon, many senior staffs rode his drivable suitcase around the event’s entrance. I wanna give this AirWheel as one of the MVP of the event. 🤣
In the few minutes later, our lovely Microsoft MVP, P’Thor Boonthawee joining the morning conversation with us, along with P’Kom, P’Ploy, P’Jimmy, and P’Vin Vasupon (Azure Manager). P’Thor has shared some experience of driving on the way coming to QSNCC with clear roadway on Labour’s Day. After that, it’s time to move ourselves to the booth inside the event area, to be ready for making a presentation of Microsoft MVP booth.
After we arrived to the booth, we warmly welcomed Aj. Bond Surasuk (Infinitas) and Aj. Komes (KBTG) to the booth for this early morning. We were preparing the setup even on slides and tables for letting visitors visiting our booth in the morning before the event opened. After we taking a preparation for a while, P’ Phol Teerasej were joining with us for a set up, along with P’Ponggun (T.T. Software) and P’Chalaivate (9Expert). Finally, we decided to group up and have a photo group of Microsoft MVP and Microsoft Learn Student Ambassadors.
8 a.m. — Time to call the visitors entering the event
We could witness that the first group of visitors entering the event was entering at 7:45, right after we finished setup on our booth. One of the person that came along with first group of visitors was P’ Ken Nakarin Nawakijpaibul (CEO and Editor-in-Chief of The Standard). We kinda had some little conversation together about the event thing and his book name “The Invisible Leader” that he wrote for a long time. Sadly, I kinda busy for connecting with many visitors around here, so that I said to him that I will talk to him in later time, he agreed and continued visiting P’Job Jirachai booth after that.
I decided to walk around the event to have a connection with many people around there and get the event’s wristband, one couple that I had met them before in the event of Global AI bootcamp, was coming in MSBuildAIday and joining with me on the queue of receiving wristband, which are P’Aim Khanittha and P’Poom Pakpoom from SCG. Lately, I got a phone call from my favorite professor, Aj. Kaew (Ph.D. Sansiri Tarnpradab), that she would join the event in real quick to meet me, along with her friend, Ph.D. Pat Khumprom from GMI KMUTT.
In the later minutes, I inspected the coming of the President of Big Data Institute, Aj. Pom Tiranee Achalakul, our former professor of CPE KMUTT. Surprisingly, our CPE family members, including me, P’Kom, P’Ploy, and Aj. Kaew coming to Aj. Pom at the same time, we decided to have a chitchat together about how the changes of CPE and each other introduced to Aj. Pom for making a connection, and having a photo group together. For now, I don't know what the name of the event really is…Is it MS Build AI Day or CPE KMUTT World Tech Forum?
8:55 am. — Our Village Leader’s COMING! + Dream came true!
I saw a lot of people were grouping together at the hallway entrance, and I didn’t realize what was happening at the moment. But I kinda guessing at the first chance that Satya was coming on that way, so I call all MVPs to come on the entrance, waiting Satya to come outside the from the exhibition hall. And now, one of the greatest moment of my lifetime has begun.
At 9 am sharp, our village leader Satya Nadella coming out from the exhibition hall, and being ready to take a photo with the group of the inspiring tech developers, with having K. Dhanawat (our MD) at the right side. However, Aj. Bond and I tried to take a selfie shot with Satya…and this was the result that we could get from him. 😭
By the way, good luck for me that I take the group photo with Nadella in them, and I don’t know that I can call “luck” or not since Nadella accidentally look at my camera at the moment.
After, he entering the exhibition hall again, the group of people is separated each other to go somewhere, some of them waiting at the entrance of the stage hallroom, some were still connecting each other since the door of the stage hallroom was not opening yet. Then, we could catch up with P’Sarah (ThaiCySec), along with our community members, and have a little talk for asking her moment on taking a group photo with Satya Nadella.
9:15–9:45 am. — Ready to be in the hallroom
At 9:15, it’s time for letting us coming in the stage hallroom. A thousand of visitors tried to find for best seats (actually front seats) to see the opening panel session from Satya Nadella. Luckily, I really got a best seat for me, which was situated at the aisle of the fifth row, and realized that Aj. Kaew and Aj. Pat was behind me, sitting together at the aisle. Wait a minute…is that P’Ponggun at the seventh row? 😭
Somehow, I tried to find something to do on waiting Satya entering the hallroom (and tried to not run my phone out of battery since I have only 90% at that moment, and I needed to take a lot of photos in the event ^o^). I just rotated my head and found out that Aj. Bond and P’Phol sitting across from us (including Aj. Kaew and Aj. Pat).
And now the greatest opening session from Satya Nadella was going to be started…NOW!
Address and Keynote by Satya Nadella
(I hope you appreciate my photos since I sat at the aisle which is very near to the front row, so I can see him very clearly). If you want to see more, please visit my photo album here >>> https://photos.app.goo.gl/oYkuXytzhwhq4SaB9
I ensure that a lot of people are talking a lot about “having our first datacenter in Thailand”, so I’m not gonna talk much about that in details.
- In this keynote, Satya has coming up with the topic of “Age of AI”, which becomes more powerful than before, which consist with three main things that would combine it to be the “innovation”: Universal interface for users, Reasoning & Planning for models, and Memory to storing the data.
- After the discussion, he came up with some statistical thing for our Thailand GDP growth, which will be reached to $117B++ if we integrate with the use of AI by 2030. This would be the vision from Microsoft that would like to “empower” every person and every organization in Thailand (mentioning National Health Security Office-NHSO, PTT Global Chemical, SCBx, Agoda, and especially AIS), and on the planet to achieve more, by the term of productivity and wellbeing.
Another big key points that Nadella mentioned was about 5 main pillars that will empower our Thailand to have more skill and growth in Technologies in the term of “Copilot Stack + AI safety and Security”, which helps you orchestrate your assistant (even with Microsoft Copilot or your own built copilot): Microsoft Cloud, AI infrastructure, Foundation Models, Your Data, and AI toolchain.
1. Microsoft Cloud
- Microsoft Cloud becomes one big group of innovations that productively do anything that helps you work easily, which contains the use for data&AI, Digital & app innovation, business app, security, and more. In meantime of having AI transformation, it will enrich employee experiences reinvent customer engagement, reshape biz processes, and bend the curve on innovation.
- By the support of Microsoft technologies, we can make our task more productive with three main things: Microsoft Copilot, AI solutions from Microsoft Azure, and Security that protect your data for your business.
- The use cases that he mentioned was from Office of the Council of state and AIS (Telecommunication).
- HIGHLIGHT OF THE EVENT: In nowadays, Microsoft Cloud (empowered with Microsoft Azure) have 60+ supported Azure Regions with 300+ Datacenters worldwide. And yes! Thailand is going to be one of them, as having our first datacenter.
2. AI infrastructure
- Satya declared us that we have the BEST AI infrastructure with the performance of the Azure NC H100 v5-series on GPT-J 6B, 11.6x faster than others when running with servers and 8.6x faster than other when it’s offline.
3. Foundation models
- In Azure OpenAI and Azure AI Services, you can use many models as you want within the service for your best selection of frontier, and it’s open-sourced (some of them).
- HOWEVER, I would like to introduce you the newest model from Microsoft “Phi-3”, a family of open AI models developed by Microsoft. This 3.8B LLM are the most capable and cost-effective small language models (SLMs) available, outperforming models of the same size and next size up across a variety of language, reasoning, coding, and math benchmarks. And now, it’s available on Azure AI services, Hugging Face, and Ollama. Also, this model will help you in the part of language, reasoning, coding, and math benchmarks (even on Chemistry, Physics, and Life sciences). Containing the power of cloud supercomputer (HPC + AI + Quantum)
4. Your data and AI toolchain
We could ensure with “AI Safety + Responsible AI” guidelines that Microsoft gives us a promise, and also adding up with security in it, to make sure that the data is not used to train or enrich foundation AI models, and covering Customer Copyright Commitment.
Apart from these pillars would be the demo GitHub Copilot Workspace, that you can use GitHub Copilot from your web-based GitHub, and it will automatically guide you on what things you need to do as your desires. And the good news is, GitHub Copilot (with Copilot Chat) supports Thai Language.
After all, it’s such a wonderful opening keynote that I’ve never seen before, such an epic moment once in my lifetime to learn new AI technological things from CEO of Microsoft like Satya Nadella and GitHub’s Sr. Manger Karan M.V. And we ended with video of inspiring tech person in the part of Skills for Jobs, with P’ Jidapa Nitiwirakun.
Story of P’Jidapa
Jidapa Nitiwirakun, a remarkable individual, faced early challenges when she was diagnosed with muscular dystrophy. Despite these obstacles, she demonstrated incredible resilience and determination. Jidapa attended Pattaya Redemptorist Technological College for People with Disabilities, where she honed her skills and participated in Microsoft’s Skills for Jobs program. Through this program, she gained valuable digital skills related to business and artificial intelligence.
Jidapa’s journey didn’t end there. She secured a job at Toyota Tsusho’s Thai headquarters, working remotely from her hometown of Pattaya. Her role in the HR department allows her to contribute to projects using AI tools. Beyond her professional achievements, Jidapa lives independently and supports her family financially. Her personal triumphs serve as an inspiration to others, showcasing the power of determination and adaptability.
Jidapa Nitiwirakun’s story reminds us that challenges can be overcome with the right mindset and support. Her positive outlook and commitment to learning have propelled her forward, proving that disability is not a barrier to success.
Session Resources
- Thailand Visit 2024 (microsoft.com)
- Microsoft announces significant commitments to enable a cloud and AI-powered future for Thailand — Microsoft Stories Asia
- Introducing Phi-3: Redefining what’s possible with SLMs | Microsoft Azure Blog
- Equipped with AI and technology skills, women across Southeast Asia find new career opportunities — Source Asia (microsoft.com)
- Microsoft announces AI skilling opportunities for 2.5 million people in the ASEAN region by 2025 — Microsoft Stories Asia
Code; Without Barriers: Inclusion is Innovation
In this panel discussion, we heard three different people are talking on making diversity of gender of their workspace and inside technology, even on big data or training for artificial intelligence. One of them who is going to give the discussion is our Aj. Pom from Big Data Institute, following with Dr. Arak and P’Gatuk. Let’s see what they were talking.
From Aj. Pom Tiranee Achalakul
She spoke about the high state silos and the need to break down these state units to facilitate data sharing. Solutions such as ‘coploit’ are to be reinforced to achieve this. BDi has a diverse team, with women making up 40% of the workforce. Moreover, the team comprises individuals from various educational backgrounds, such as Mechanical Engineering and Geography, not just Computer Science. This diversity allows for effective coordination with different units and a better understanding of the domain. The age range of the team is between 22 and 55 years.
She emphasized that in AI, one cannot just throw models; user experience must be prioritized. Understanding the user will enable the development of suitable AI and yield benefits. When it comes to datasets, if there is diversity in the people labeling the data, it can help reduce bias.
From Dr. Arak Sutivong(SCBx)
He spoke about AI and how it used to be a distant concept until the advent of ChatGPT, which brought it closer to us. SCBx has transitioned into an AI-first organization, leveraging AI to assist in various areas, generate revenue, reduce costs, and enhance user experience.
However, given the diversity of the workforce, becoming an AI-first organization requires building up people’s readiness. It’s crucial to understand what AI is before diving into it. This understanding is divided into three groups:
- Foundation: Understanding what AI is, how it’s used, and the dos and don’ts of responsible AI.
- Intermediate: Utilizing low code.
- Advanced: Involving coding for development and data science.
Once employees understand these aspects, they can start adapting, speeding up their work, and increasing opportunities.
From P’Gatuk Sudarat Chattanon
She spoke about PyLadies, a program designed for women to start coding. In the past, there might have been barriers to learning, such as the need to open books and enroll in classes. Now, AI has stepped in to assist, with tools like GitHub Copilot.
The more people get involved, the more it helps drive the growth of tech in Thailand. Having programs like this encourages women to start coming together. In the past, the tech field was predominantly male, which could make it challenging for women to get started.
However, there are also obstacles, especially during the COVID period, which has led to significant behavioral changes, such as shifting from offline to live. This transition can result in a lack of activities. To address this, tools are provided to encourage people to communicate more and reduce shyness.
Session Resources
- Code; Without Barriers — Asia Website (microsoft.com)
- (7) Code; Without Barriers | Groups | LinkedIn
- Microsoft Code; Without Barriers Hackathon 2024: Empowering Women in AI — Code Without Barriers Hackathon 2024 (cwbhack.dev)
- Asia AI Odyssey from Microsoft
What’s New in Generative AI
P’Jimmy Danai and P’Si Mullika discussed the current focus on ubiquity, multi-modal (capable of handling various formats beyond text, images, and sound), and autonomy. They suggested adopting tools like Copilot, extending Copilot with low-code/no-code solutions, and building your own solutions using Azure AI and Microsoft Copilot Studio.
They mentioned the new Azure AI Studio tool, which helps manage the flow of LLMOps and AI Safety. It offers features like Explorer for model exploration, Build for trying and building, and Manage for handling different permissions.
They also talked about the new ‘prompt flow’ feature, which introduced in the Global Azure session from P’Phol Teerasej, which helps capture data from the developer flow, such as UI. The prompt flow is seen as an endpoint for developers to ask and answer questions, similar to a BPMN Engine, but specifically for AI.
They highlighted the capabilities of GPT4 Turbo with Vision Model for image/video analysis. For instance, it can explain a graph, describe initial damage from an insurance claim image, or explain a video in a chat.
They also discussed building your own custom neural voice from the speech studio. This allows you to create an assistant with your voice that can speak in different languages. In addition to voice, they mentioned the use of avatars in live chat (Azure AI Avatar Studio).
Used Case Scenario — SCBx
P’Yada Sareesavetrat, Senior Consumer Insights and Market Research Expert at SCBX, along with P’Nut Chukamphang, Senior AI Scientists at DataX, discussed a system that alerts when unsatisfied comments abour banking system occurs on social media. This is important because such incidents can reduce credibility, potentially affecting revenue, value, currency, and stock prices. Previously, a team would monitor these situations, but now AI is used to assist, specifically Azure Open AI.
Typically, it occurs in three stages:
- Genesis of Incident: The initial stage where the goal is to deploy AI.
- First Surge: The stage where people start searching or retweeting, causing hashtags to trend.
- After Shock: This stage often involves some drama and frequently appearing hashtags.
In the past, human intervention was required at stage 2, which often led to widespread effects. The AI used has the following architecture:
- Semimetal Analysis + Insight: Analyzes the drama that occurs, whether it’s positive or negative for any company in the group.
- Incident Detection and Summarization: Uses GPT to summarize what happened.
- Mitigation Plan: Provides recommendations for people to read and make decisions. Ultimately, a person must make the final decision.
- Dashboard: Summarizes where and what happened.
By using AI in this way, it helps increase cost efficiency and provides 24/7 monitoring for brand reputation protection. The next step is to scale and improve accuracy.
Session Resources
- Copilot for Microsoft 365 — Microsoft Adoption
- What is AI Studio? — Azure AI Studio | Microsoft Learn
- Azure OpenAI Service — Advanced Language Models | Microsoft Azure
- Prompt flow in Azure AI Studio — Azure AI Studio | Microsoft Learn
Vector search and state-of-the-art retrieval for generative AI apps
In the session, P’Smith and P’Aouychai discussed Vector Search via Azure AI Search. He highlighted the limitations of the LLM model, such as outdated or internal knowledge management (KM), which includes organizational cases. To address these limitations, P’Smith proposed three solutions:
- Prompt Engineering: In context learning
- Fine Tuning: This involves training the model to acquire new knowledge.
- RAG: This method involves summarizing the information from the documents.
In Vector Database: This solution addresses the problem of identifying similar groups of words, such as “สุนัข” and “Golden Retriever”, or “ข้าวต้มกุ้ง” and “ข้าต้มกุ้ง”. The solution involves viewing these as vectors and using mathematics to determine their similarity or proximity using algorithms such as cosine similarity. Once the vectors are obtained, they are stored in a database and indexed.
P’Smith and P’Aouychai also discussed various Vector Search Techniques:
- ANN: This is a fast method.
- Exhaustive ANN: This method emphasizes accuracy.
- Filter Vector: This reduces the scope context to speed up the search, in addition to using ANN.
- Multi-Vector: This method is used to address the challenge of long documents reducing accuracy.
To solve this, P’Smith and P’Aouychai suggested us using Multi Model Azure, which includes Fusion (RRF(Reciprocal Rank Fusion) keyword + vector) and Reranking (aligning with the domain based on semantics, such as financial documents or food) to supplement keywords or vectors.
Session Resources
- RAG and generative AI — Azure AI Search | Microsoft Learn
- azure-search-vector-samples/demo-python at main · Azure/azure-search-vector-samples (github.com)
- Vector search — Azure AI Search | Microsoft Learn
Lunch Time + Networking
After I finished attending morning session and gave a little bit take-caring Microsoft MVP booth (and meeting the team of BorntoDev, even P’Prem, P’Aeff, and P’New), I decided to get something to eat for my lunch. Luckily, QSNCC have a very luxurious food box with eco-friendly used materials for containing food and desserts in it.
Let’s see the box set
Here is my box menu list (no pork and no lard):
- Wok Fried crab in yellow curry
- Roasted BBQ Duck served with Bok Choy
- Sear Australian grilled sirloin (didn’t mean to eat since I have Goddess Guanyin’s blessing practice😭🙏)
- Fresh Rice Noodle Roll with Prawn and Minced Chicken
- Deep-fried fish cake served with Sweet Chili Sauce
- Orange juice
- Some kind of Muffin?
Meeting more people
During lunch time, I had met some similar faces that I’ve met them before, even from CreatorsGarten (like P’RayRiffy and P’Pub), P’Earn from Code Without Barriers, P’Got (padagot-MSP), and my blog collab like P’Ping Chatri (Naiwaen debuggingsoft), sharing each other’s experiences and blog notes from attending the morning session, and also sharing Satya’s photos together.
Meeting Seven Peaks’ Partners
Lately, it’s been a very great time to have a conversation with people from Seven Peaks, the company where I take much time on contributing and collaborating with community in the part of Azure and .NET framework. One newest person that I haven’t had a conversation with him before is Seven Peak’s CEO Jostein Aksnes, he shared some experience on the morning session with me, and later Jose Barbosa is joining the chat.
Build your RAG Application with Prompt flow in Azure AI Studio
By this session, P’Jiratouch combined the RAG (Retrieval-Augmented Generation) approach with Prompt Flow. The LLMOps Flow involves three phases for continuous improvement:
- Ideating / Exporting Use Cases: This phase focuses on identifying use cases for AI, filtering out irrelevant ideas, and piloting the selected ideas.
- Building / Augmenting: In this phase, real data is used to build and evaluate models, including checking for sensitive information.
- Operationalizing: After going live, monitoring and adjustments take place.
Concerns related to developing applications using Generative AI include:
- Private Data Access / Data Privacy
- Prompt Engineering: Techniques like “Chain of Thought” and “Few Shot.”
- Performance & Evaluation
- Versioning: Similar to managing code versions.
- AI Safety
Regarding Prompt Flow, when deploying, it configures metrics for AI safety. If you’re developing Prompt Flow based on what you heard from the Session Global Azure 2024, there are two approaches:
- Designer First: Web-based flow design.
- Code First: Use the VS Code extension for Prompt Flow and Python runtime. You can start locally before expanding to the cloud by adjusting the flow via YAML.
Used Case Scenario — Agoda
Since K.Shaun participating in the GPT Hackathon, we can explore various examples, such as chatbots, SEO content generation, and hotel price recommendations. To bootstrap a community, they open up knowledge sharing to everyone, gradually expanding AI adoption across all roles. Developers are encouraged to try Copilot, similar to what P’Jimmy and P’Si discussed during the morning session. Additionally, within an organization, they can build our own AI tools, including:
- AgodaGPT
- Assistance Tools
- AI-Generated UI Components (from Drawings to React)
- Query Assistant for SQL Generation from Organization Databases
- Askgoda (Semantic Search)
- AI Review Helper
- Property Page Q&A Chatbot
Looking ahead, the future involves integrating AI directly into applications. For instance, if we inquire about hotels, the system will not only recommend but also provide reasoning behind its suggestions. This could include chat flows, assistance, and more.
Session Resources
- Get started with RAG using a prompt flow sample (preview) - Azure Machine Learning | Microsoft Learn
- Build a RAG-based copilot solution with your own data using Azure AI Studio (preview) — Training | Microsoft Learn
Prompt Engineering to Production with LLMOps
This session is viewed as a continuation of the previous session, focusing on the building and augmenting aspects. Here were some information that P’Lin, P’Phol, and Aj.Bond discussing within this session.
ML vs LLM
- In ML, you create your own model.
- In LLM, you take an existing model and enhance it with data to fine-tune it, making it smarter, plus the prompt.
- Discussing MLOps and LLMOps.
Then there’s a demo for “Contoso Outdoor Company” with two parts: Dev Flow and Evaluate Flow.
- Dev Flow: Similar to other sessions, we recommend looking at the flow and components live. If you can write yaml GitHub Action, it’s not difficult to adjust.
- Evaluate Flow: Write another Prompt Flow to check steps, define a json file to insert prompts for each intent, test locally and if ready, go to the cloud.
Metrics used during evaluation:
- Groundedness: Is the answer correct? Does it have supporting facts or is the AI making it up?
- Relevance: Does it match the requirement? It shouldn’t be pulled into another category.
- Coherence: Does the answer seem human-like?
Next, we demo the Evaluate Flow. The flow’s appearance separates the metrics for dragging and dropping, and defines jsonl to perform two tasks: Assistance and Support. The result shows that the metrics are not pretty. Then adjust the Prompt Flow, do Step Intent Mapping to help choose Assistance more accurately whether to trigger Assistance or Support, and adjust Prompt Normalization to be clearer / Rewrite Query.
Tooling for Automation
- Azure CLI / Azure ML CLI / PF Azure CLI
- Prompt Flow yaml
- GitHub Action needs to be configured.
Session Resources
- Elevate Your LLM Applications to Production via LLMOps — Microsoft Community Hub
- The new AI imperative: Unlock repeatable value for your organization with LLMOps | Microsoft Azure Blog
Make your data AI ready with Microsoft Fabric
I ensure that many of us may know a little bit information about Microsoft Fabric before this event, like on community meetup, or small events. However, by this event, there are some new updates from Microsoft. So that, in this session, P’Ple, P’Zeng, and P’Job, would like to update something news from the updated version of Microsoft Fabric in this event.
Firstly, Microsoft Fabric is a complete analytic platform that works across seven domains:
- Data Factory: Handles flows for ETL (Extract, Transform, Load) and ELT (Extract, Load, Transform) processes. It connects to over 150 data sources.
- Synapse Data Engineering: Manages data through Apache Spark for data analytics.
- Synapse Data Science: Analyzes data and trains models.
- Synapse Data Warehousing: Manages large-scale data for OLTP (Online Transaction Processing).
- Synapse Real-Time Data Analytics: Processes real-time data.
- Power BI: Displays data visualizations.
- Data Activator: Generates alerts from data.
Lake Centric — Onelake:
- Speaking about Onelake, it consolidates data from various sources into a single location. It operates at the Onelake tenant level and includes Purview for data governance.
- Notably, it prevents data duplication and links data from other storage solutions like Azure Data Lake, S3, and Data Brick.
- The Mirroring feature in Microsoft Fabric (currently in Public Preview) allows data to be pulled from different databases, such as Azure SQL Database, Azure Cosmos DB, and Snowflake, into a unified system.
Empower Every Business User:
- Integrates with other tools like Teams, PowerPoint, and Excel. Users can seamlessly utilize the Lake House concept.
AI Power — Copilot:
- Assists in making data ready within data flows by performing transformations.
- In notebooks, it suggests code for data visualization and forecasting.
- Helps discover insights and create reports in Power BI
I ensure that Microsoft Fabric is currently available in public, and accessible through organizations and enterprises. If you want to know more about the updated version’s information, please visit at the session resources below.
Session Resources
- Microsoft Fabric release plan documentation — Microsoft Fabric | Microsoft Learn
- Fabric-Readiness/events/2024-microsoft-ai-tour.md at main · microsoft/Fabric-Readiness (github.com)
Post-Event + Networking
Being the same stage with Satya Nedella would be great at that time, so that even CreatorsGarten, P’Job (again), and me, would like to get some photos as an event souvenir. We had photos’ name called “The Hidden Session”, the session which didn’t happen in the event. We just did it for fun somehow.
After that, we have a very last group photo together after finishing Microsoft Fabric session and it’s time to say goodbye for a very big stage hallroom. And then, we leave from the hallroom and returned to Microsoft MVP Booth.
CPE KMUTT Meetup
After we got out from the main hallroom, we still saw at that moment which some of us still in Hackathon room so far. Although, we did meet and made some reunion of CPE KMUTT family members who attended the event today. Aj. Kaew and I decided to call our seniors and juniors taking a photo together and then we are going to leave each other, saying goodbye and back home.
Meeting with MD, along with Microsoft Senior Staffs
In the final section of staying in the exhbition area, many Microsoft senior staffs gathered to make a photo group together with the Microsoft logo statue. After that, I kinda have nothing to do except have a little chitchat among seniors. Additionally, I did meet K.Dhanawat, K.Chanikarn and P’Vin Vasupon, have some conversation and making a connection together via LinkedIn. And then, we have take a photo together (as also my souvenir from the event…again🤣).
In conclusion,
I wanna say that Microsoft Build: AI Day Thailand 2024 is the cumulation for summary of everything that has been through over 2 years of the Age of AI, also with learning new innovations that lately introduced to us like Phi-3 library, Promptflow, GitHub Copilot Workspace, and Azure AI RAG Vector Search. We did have a very great moment so far, having laugh and smile even with Satya Nadella, many executives, many visitors, along with our senior staffs from Microsoft Thailand.
By the name of Microsoft Student Ambassadors, I would like to give an appreciation to Satya Nedella, P’Toey Dhanawat Suthumpun, P’Noi Chanikarn Pronanun, along with all Microsoft Thailand staffs (including P’Chang Phantip — our community leader) for having me along with fellow MLSA and visitors in this biggest event of my career. By meantime, I would like to thank you to Pablo Veramendi and Tracy Salem for giving me an opportunity to be in the part of the organization by the name of MLSA.
Finally, Microsoft Thailand has been with us for over 31 years, and still counting. We hope that all clients will have a very good experience on using our products and services always and along to the future, along with sharing some new innovations through social media and blogs. We hope that this MegaBlog would be benefit for all of you more or less, and hope that you guys gonna like this MegaBlog.
Now, it’s time to go home with P’Chang and P’Krit. See you in the next blog of Python Developer Meetup (11th May) and following with GitHub Galaxy (17th May). Have a good time guys. 👋
