032. PyThaiNLP - Open source tool for Thai Natural Language Processing

We are back! For this episode, we discussed topics about generative AI. We actually used Google's Bard to generate the outline for this episode. It covers topics: What is generative AI? How does it work? What are some of the benefits of generative AI? Challenges of generative AI, the future of generative AI, and many more! Overall, this episode provides a comprehensive overview of this rapidly developing field. It could also explore the ethical implications of generative AI, and its potential impacts on society. Let us know what you think. You can reach us directly from our LinkedIn page. Koo Ping Shung: https://www.linkedin.com/in/koopingshung/ Thu Ya Kyaw: https://www.linkedin.com/in/thuyakyaw/

It has been a while since we released an episode on this channel. Apologies, we were both busy with work and couldn't find a common dedicated time to record an episode. We also changed the intro and outro music. Let us know what you think? In this episode, Koo Ping Shung and I discussed the nitty gritty things about AI ethics. We also voiced our opinions on the different aspects of AI ethics and whether having a governing body to control the ethics aspect of AI is a good thing or not. Have a listen! If you have any feedback, you can send them our way from here: https://forms.gle/cdgUUtsmdnsrNUPMA Want to appear in our podcast episode? Let us know from here: https://forms.gle/g9xoC12eEUSA6vhdA

This week, we invited Kelvin Tham, an MLOps Data Program Manager at ViSenze - AI for Visual Commerce. Kelvin has a wide range experience across ML Ops, data analytics, and business process improvement. He is currently working on design, development and shipping of ML Ops model management. In this episode, he shared about how is it like to be working at an AI startup company and his war stories of wearing multiple hats at one go. He also talked about the differences between being a program manager and a developer, pros and cons of each role, and shed a light on what to look out for when you are exploring your future career options. Have a listen. You can connect with Kelvin here: https://www.linkedin.com/in/kelvinthamkh/

This week, we invited Teck Liang Tan (PhD), a Senior Data Scientist @ NTUC Enterprise to have chat with us. He walked us through his unconventional career move from Physics to Data Science. Also, his reason of getting into the industry instead of continuing in academia. By the way, do you know what is complexity science?  If you are curious, you should definitely give this episode a listen! He also shared learning resources for getting into the field as well as keeping the skills sharp. We also talked about #MajulaGCP season 5 which is happening right now and many more! Teck Liang's profile: https://www.linkedin.com/in/teck-liang-tan-47a47327

This week we invited Ivan who is a Lead Data Scientist at Tech in Asia. Ivan has a unique set of technical competencies, project management, interpersonal skills and problem solving abilities. He is also experienced in deploying scalable machine learning systems, data engineering pipelines, dashboards and delivering actionable insights through the use of statistics and data visualization. In this episode, Ivan shared his career journey from being an undergraduate to leading a data science team. He also shared whether doing an internship is useful and many more interesting tips and tricks to land a job in the data industry. Check this out!

Data collection is a crucial step for any data related projects. So much so that you might have encountered something along the lines of the “GIGO” (garbage in, garbage out) concept. Some might even say having the right data is more important than having tons of data that can’t be used. As web scraping being one of the ways to collect data, for this episode, we invited Cliff, a data consultant, back to discuss his personal experience with web scraping. He shared topics such as the basics of web scraping, web scraping tools, the challenges that he faced while trying to scrape web contents, ethics of web scraping, learning materials, and more! Resources: Cliff's medium post 1: https://medium.com/codex/scraping-singapore-libraries-f74c541f1f94 Cliff's medium post 2: https://cliffy-gardens.medium.com/iterations-for-my-nlb-scraper-github-code-provided-b4e1f1bd422e Selenium: https://www.selenium.dev/ BeautifulSoup: https://www.crummy.com/software/BeautifulSoup/bs4/doc/ TagUI: https://github.com/kelaberetiv/TagUI Web Scraping with Python: https://www.oreilly.com/library/view/web-scraping-with/9781491985564/

For this episode, we invited Low Yi Xiang, a Data Scientist at Traveloka again to have a chat about Data Product Management. Yi Xiang covered what is data product management in a nutshell, how it differs from the other product management practices, what are the stages of data product management lifecycle, and many more interesting topics. We hope you enjoy this episode as much as we had fun recording and producing it. Yi Xiang is an experienced data scientist who likes to work on various sorts of data problems. He also possesses a strong track record of being a strong individual contributor and/or taking on lead positions delivering huge impact. Although his title is data scientist, he works on a wide range of scopes, whether it is analytics, data engineering, building models and moving models to production and post monitoring. If you want to catch up with Yi Xiang, he can be reached via his LinkedIn: https://www.linkedin.com/in/yi-xiang-low-b349137b/

In this episode, we invited one of our popular guests, Charin Polpanumas, back! We got him to share a project that he is passionate about, PyThaiNLP. In this episode, we discuss the challenges of Natural Language Processing and also creating and working on an open-source project. This episode is definitely for anyone who is interested in Natural Language Processing as we discuss many aspects of NLP, building corpus, challenges in translation, and challenges on the limited training datasets! Do check it out if you are someone passionate about NLP! Reference Source: https://github.com/PyThaiNLP/pythainlp

So what is MLOps? This is a topic we covered in this episode. We discuss the different aspects of MLOps, for instance, data, business requirements, and also measuring the performance metrics. We discuss also data quality and feature engineering and its impact on the ML pipelines as well. We also do a short introduction on the different tools used in MLOps, such as Containers, Kubernetes, and Airflow. And let us throw in one more technical term...data versioning. Give us a listen to understand what that is! Learning Resources: 1. What is MLOps (https://whatis.techtarget.com/definition/machine-learning-operations-MLOps) 2. Getting started with MLOps (https://ml-ops.org/) 3. MLOps Fundamentals with GCP (https://www.coursera.org/learn/mlops-fundamentals) 4. Difference between Data Scientist and MLOps Engineer (https://towardsdatascience.com/data-scientist-vs-machine-learning-ops-engineer-heres-the-difference-ad976936e651) 5. Learn Docker (https://www.youtube.com/watch?v=fqMOX6JJhGo) 6. Learn Kubernetes (https://kubernetes.io/docs/tutorials/kubernetes-basics/) 8. https://www.deeplearning.ai/program/machine-learning-engineering-for-production-mlops/

In this episode, we have another guest - Amelia, a chemical engineer turned data scientist! Listen to the episode to understand more about her successful transition, what are the skills that she finds valuable as a data scientist, and how did she cope with studying for Master's and working at the same time. We had a great discussion on the topic of coping with work, studies, and everything else! If you want some tips and tricks, tune in to this episode to find out more! Amelia also shared about her Global Health Fellowship experience with an NGO in Seattle, and how she used her data science skills for a better world! It was an eye-opening experience that taught her about change management and how to ensure the data science momentum persists in an organization. Amelia's LinkedIn Profile: https://www.linkedin.com/in/pehyingqi/