Podcast appearances and mentions of chris albon

MLOps Coffee Sessions #68 with Chris Albon, Wikimedia MLOps co-hosted by Neal Lathia. // Abstract // Bio Chris spent over a decade applying statistical learning, artificial intelligence, and software engineering to political, social, and humanitarian efforts. He is the Director of Machine Learning at the Wikimedia Foundation. Previously, Chris was the Director of Data Science at Devoted Health, Director of Data Science at the Kenyan startup BRCK, cofounded the AI startup Yonder, created the data science podcast Partially Derivative, was the Director of Data Science at the humanitarian non-profit Ushahidi, and was the director of the low-resource technology governance project at FrontlineSMS. Chris also wrote Machine Learning For Python Cookbook (O'Reilly 2018) and created Machine Learning Flashcards. Chris earned a Ph.D. in Political Science from the University of California, Davis researching the quantitative impact of civil wars on health care systems. He earned a B.A. from the University of Miami, where he triple majored in political science, international studies, and religious studies. // Relevant Links --------------- ✌️Connect With Us ✌️ ------------- Join our slack community: https://go.mlops.community/slack Follow us on Twitter: @mlopscommunity Sign up for the next meetup: https://go.mlops.community/register Catch all episodes, Feature Store, Machine Learning Monitoring and Blogs: https://mlops.community/ Connect with Demetrios on LinkedIn: https://www.linkedin.com/in/dpbrinkm/ Connect with Neal on LinkedIn: https://www.linkedin.com/in/nlathia/ Connect with Chris on LinkedIn: https://www.linkedin.com/in/chrisralbon/

In this episode we're joined by Chris Albon, Director of Machine Learning at the Wikimedia Foundation. Lukas and Chris talk about Wikimedia's approach to content moderation, what it's like to work in a place so transparent that even internal chats are public, how Wikimedia uses machine learning (spoiler: they do a lot of models to help editors), and why they're switching to Kubeflow and Docker. Chris also shares how his focus on outcomes has shaped his career and his approach to technical interviews. Show notes: http://wandb.me/gd-chris-albon --- Connect with Chris: - Twitter: https://twitter.com/chrisalbon - Website: https://chrisalbon.com/ --- Timestamps: 0:00 Intro 1:08 How Wikimedia approaches moderation 9:55 Working in the open and embracing humility 16:08 Going down Wikipedia rabbit holes 20:03 How Wikimedia uses machine learning 27:38 Wikimedia's ML infrastructure 42:56 How Chris got into machine learning 46:43 Machine Learning Flashcards and technical interviews 52:10 Low-power models and MLOps 55:58 Outro

Sponsored by us! Support our work through: Our courses at Talk Python Training Test & Code Podcast Brian #1: Building a self-updating profile README for GitHub Simon Willison, co-createor of Django “GitHub quietly released a new feature at some point in the past few days: profile READMEs. Create a repository with the same name as your GitHub account (in my case that’s github.com/simonw/simonw), add a README.md to it and GitHub will render the contents at the top of your personal profile page—for me that’s github.com/simonw” Simon takes it one further, and uses GitHub actions to keep the README up to date. Uses Python to: Grab recent releases from certain GH repos using GH GraphQL API Links to blog entries using feedparser Retrieve latest links using SQL queries Michael #2: Handcalcs Created by Connor Ferster In design engineering, you need to do lots of calculations and have those calculation sheets be kept as legal records as part of the project's design history. If they are not being done by hand, then often Excel is used but formatting calculations in Excel is time consuming and a maintenance nightmare. However, doing calculations in Jupyter is not any better even if you fill it up with print() statements and print to PDF: it just looks like a bunch of code output. Even proprietary software like MathCAD cannot render math as good as a hand calculation because it does not show the numerical substitution step. No software does Why handcalcs exists: Type the formula once into a Jupyter cell Have the calculation be rendered out as beautifully as though you had written it by hand. Write your notebooks once, and use them for calculation again and again; the formula you write is the same as the representation of the formula. **Symbolic** The primary purpose of handcalcs is to render the full calculation with the numeric substitution. This allows for easy traceability and verification of the calculation. However, there may be instances when it is preferred to simply display calculations symbolically. For example, you can use the # Symbolic tag to use handcalcs as a fast way to render Latex equations symbolically. Includes longhand vs. shorthand Use units (mm^3) for example. Brian #3: The (non-)return of the Python print statement Article by Jake Edge Idea by Guido van Rossum to bring back the print statement. Short answer: not gonna happen Michael #4: FastAPI for Flask Users Flask has become the de-facto choice for API development FastAPI that has been getting a lot of community traction lately Benefits Automatic data validation documentation generation baked-in best-practices such as pydantic schemas and python typing Running “Hello World” - super similar, but FastAPI is uvicorn out of the box @app.get('/') vs @app.route('/') FastAPI defers serving to a production-ready server called uvicorn. URL Variables: Flask @app.route('/users/[HTML_REMOVED]') def get_user_details(user_id): - FastAPI @app.get('/users/{user_id}') def get_user_details(user_id: int): Query Strings Flask @app.route('/search') def search(): query = request.args.get('q') FastAPI @app.get('/search') def search(q: str): Taking inbound JSON request in FastAPI: def lower_case(json_data: Dict) Nice but if you define a Sentence model via pydantic: @app.post('/lowercase') def lower_case(sentence: Sentence): Blueprints == Routers Automatic validation via pydantic Brian #5: Tweet deleting with tweepy Chris Albon A useful and simple example of using tweepy to interact with Twitter Chris set up and shared a Python script that deletes tweets that are: older than 62 days have been liked by less than a 100 people haven’t been liked by yourself Michael #6: Clinging to memory: how Python function calls can increase your memory usage by Itamar Turner-Trauring I had Itamar on Talk Python episode 274 to discuss FIL which was recently covered. This article basically uses FIL to explore patterns for lowering memory usage within the context of a function. With simple code like this, we expected 2GB of memory usage, but we saw 3GB: - def process_data(): data = load_1GB_of_data() return modify2(modify1(data)) The problem is that first allocation: we don’t need it any more once modify1() has created the modified version. But because of the local variable data in process_data(), it is not freed from memory until process_data() returns. Solution #1: No local variable at all return modify2(modify1(load_1GB_of_data())) Solution #2: Re-use the local variable data = load_1GB_of_data() data = modify1(data) data = modify2(data) return data Solution #3: Transfer object ownership See article Extras: Michael: Pickle Use Example via Adam. I once had to work on an API that spoke to a 3rd party service that was a little unusual. That communication to the 3rd party service was over a raw socket connection, so we were responsible for crafting specifically formatted byte arrays to send to them, and we'd get specifically formatted byte arrays back which we'd then have to parse out to determine what pieces of data were in the message. The other wrinkle: that service wasn't available 24/7 but only during limited specific testing periods which had to be negotiated days in advance. We instrumented the code with a feature flag to enable pickling all received messages from that 3rd party service. Python 3.8.4 is out I'm an Arctic Code Vault Contributor over at GitHub. You might be too. Joke:

Hosts new and old gather together for this special episode of the podcast! We’ll talk about our favorite episodes of the year, the coolest things from 2019, and wrap up another great year together doing what we love! Happy Holidays to all of our listeners, and we’ll see you in the new year! Top episodes of the year GCP Podcast Episode 173: Cloud Run with Steren Giannini and Ryan Gregg podcast GCP Podcast Episode 165: Python with Dustin Ingram podcast GCP Podcast Episode 175: MongoDB with Andrew Davidson podcast GCP Podcast Episode 160: Knative with Mark Chmarny and Ville Aikas podcast GCP Podcast Episode 180: Firebase with Jen Person podcast GCP Podcast Episode 164: Node.js with Myles Borins podcast GCP Podcast Episode 174: Professional Services with Ann Wallace and Michael Wallman podcast GCP Podcast Episode 176: Human-Centered AI with Di Dang podcast GCP Podcast Episode 168: NVIDIA T4 with Ian Buck and Kari Briski podcast GCP Podcast Episode 163: Cloud SQL with Amy Krishnamohan podcast Favorite episodes of the year Mark Mirchandani’s Favorites: GCP Podcast Episode 193: Devoted Health and Data Science with Chris Albon podcast GCP Podcast Episode 177: Primer with John Bohannon podcast GCP Podcast Episode 202: Supersolid with Kami May podcast Mark Mandel’s Favorites: GCP Podcast Episode 186: Blockchain with Allen Day podcast GCP Podcast Episode 196: Phoenix Labs with Jesse Houston podcast Jon’s Favorites: GCP Podcast Episode 199: Data Visualization with Manuel Lima podcast GCP Podcast Episode 196: Phoenix Labs with Jesse Houston podcast GCP Podcast Episode 206: ML/AI with Zack Akil podcast GCP Podcast Episode 201: FACEIT with Maria Laura Scuri podcast Gabi’s Favorites: GCP Podcast Episode 199: Data Visualization with Manuel Lima podcast GCP Podcast Episode 167: World Pi Day with Emma Haruka Iwao podcast GCP Podcast Episode 206: ML/AI with Zack Akil podcast GCP Podcast Episode 198: SeMI Technologies with Laura Ham podcast Favorite things of the year Mark Mirchandani’s Favorites: Cloud Run Mark Mandel’s Favorites: Stadia Samurai Shodown available on Stadia All the new podcast hosts! Jon’s Favorites: First time doing the podcast at NEXT and it was quite the experience. Going to Nvidia offices to do an episode Getting to talk to guests in the gaming industry and hear how passionate they are about the things they are building Joining the podcast Podcast outtakes! Gabi’s Favorites: Visited a bunch of offices! Joining the podcast Cloud NEXT talk, where my demo failed but I recovered! Spreading the love and joy of databases Where can you find us next? Mark Mirch’ will be sleeping as much as possible! Mandel will be working on plans for Next, GDC, and I/O 2020! Gabi will be running away to warm weather for her winter vacation! Jon will be home! He’ll also be planning gaming content for next year and wrapping up this year with some deep dives into multiplayer games and some possible content! Sound Effects Attribution “Small Group Laugh 4, 5 & 6” by Tim.Kahn of Freesound.org “Incorrect” by RicherLandTV of Freesound.org “Correct” by Epon of Freesound.org “Fireworks 3 Bursts” by AtomWrath of Freesound.org “Jingle Romantic” by Jay_You of Freesound.org “Dark Cinematic” by Michael-DB of Freesound.org “Bossa Loop” by Reinsamba of Freesound.org

Michelle Casbon is back in the host seat with Mark Mirchandani this week as we talk data science with Devoted Health Director of Data Science, Chris Albon. Chris talks with us about what it takes to be a data scientist at Devoted Health and how Devoted Health and machine learning are advancing the healthcare field. Later, Chris talks about the future of Devoted Health and how they plan to grow. They’re hiring! At Devoted Health, they emphasize knowledge, supporting a culture of not just machine learning but people learning as well. Questions are encouraged and assumptions are discouraged in a field where a tiny mistake can change the care a person receives. Because of this, their team members not only have a strong data science background, they also learn the specific nuances of the healthcare system in America, combined with knowledge of the legal and privacy regulations in that space. How did Chris go from Political Science Ph.D. to non-profit data science wizard? Listen in to find out his storied past. Chris Albon Chris Albon is the Director of Data Science at Devoted Health, using data science and machine learning to help fix America’s health care system. Previously, he was Chief Data Scientist at the Kenyan startup BRCK, cofounded the anti-fake news company New Knowledge, created the data science podcast Partially Derivative, led the data team at the humanitarian non-profit Ushahidi’s, and was the director of the low-resource technology governance project at FrontlineSMS. Chris also wrote Machine Learning For Python Cookbook (O’Reilly 2018) and created Machine Learning Flashcards. He earned a Ph.D. in Political Science from the University of California, Davis researching the quantitative impact of civil wars on health care systems. Chris earned a B.A. from the University of Miami, where he triple majored in political science, international studies, and religious studies. Cool things of the week How Itaú Unibanco built a CI/CD pipeline for ML using Kubeflow blog Why TPUs are so high-performance BFloat16: The secret to high performance on Cloud TPUs blog TPU Codelabs site Benchmarking TPU, GPU, and CPU Platforms for Deep Learning paper Machine Learning Flashcards site Interview Devoted Health site Devoted Health is hiring! site Ushahidi site FrontlineSMS site New Knowledge site Joel Grus: Fizz Buzz in TensorFlow site Snowflake site Periscope Data site Airflow site Kubernetes site Chris Albon’s Website site Partially Derivative podcast Partially Derivative Back Episodes podcast Question of the week Chris Albon To paraphrase: A computer program is said to learn if its performance at specific tasks improves with experience. To find out more, including the definition of a partial derivative, buy a pack of Chris’s flashcards. Who knows, they might help you land your next job. Where can you find us next? Michelle is planning the ML for Developers track for QCon SF on Nov. 13. Mark is staying in San Francisco and just launched two Beyond Your Bill videos: Organizing your GCP resources and Managing billing permissions. Sound Effect Attribution “Small Group Laugh 5” by Tim.Kahn of Freesound.org “Crowd Laugh” by Tom_Woysky of Freesound.org “Transformers Type SFX 2” by HykenFreak of Freesound.org “Approx 800 Laugh” by LoneMonk of Freesound.org “Bad Beep” by RicherLandTV of Freesound.org “C-ClassicalSuspense” by DuckSingle of Freesound.org

This week, Hugo speaks with Chris Albon about getting your first data science job. Chris is a Data Scientist at Devoted Health, where he uses data science and machine learning to help fix America's healthcare system. Chris is also doing a lot of hiring at Devoted and that’s why he’s so excited today to talk about how to get your first data science job. You may know Chris as co-host of the podcast Partially Derivative, from his educational resources such as his blog and machine learning flashcards or as one of the funniest data scientists on Twitter.LINKS FROM THE SHOWDATAFRAMED GUEST SUGGESTIONSDataFramed Guest Suggestions (who do you want to hear on DataFramed?)FROM THE INTERVIEWChris on TwitterChris's WebsiteDevoted WebsiteMachine Learning Flashcards (By Chris Albon)Machine Learning with Python Cookbook (By Chris Albon)FROM THE SEGMENTSGuidelines for A/B Testing (with Emily Robinson ~26:50)Guidelines for A/B Testing (By Emily Robinson)10 Guidelines for A/B Testing Slides (By Emily Robinson)Original music and sounds by The Sticks.

Chris Albon is the Chief Data Scientist of BRCK: a Kenyan startup building a network dedicated to connecting Africa to the internet, and author of the Machine Learning with Python Cookbook. For years he’s been contributing to the data science world through so many different outlets:   Cofounder of New Knowledge AI - an social media platform focused on protecting companies from disinformation, fighting fake news, and defending public discourse Former host of Partially Derivative - a popular podcast mixing explorations into data science techniques with discussions in the field’s leading experts. Content creator of Machine Learning Flashcards  - simplified, easy to digest flashcards for otherwise-complex machine learning concepts.   Blog writer at chrisalbon.com - providing some of best (and definitely most wide-ranging) technical notes out there on machine learning, statistics, deep learning, Python, and so much more. Our conversation went many places, including: How early childhood experiences (including his heritage in Zimbabwe) led him to focus his career so strictly on social impact, through political and humanitarian efforts How he’s established himself (and his blog) as an authority figure in data science & AI (with 523 resource links and counting) without having a “technical” degree The step-by-step process BRCK takes when incorporating technology into African communities & what he’s learned about challenging his own assumptions while doing so The initial inspiration behind starting the Partially Derivative podcast and why his aim was for it to be the “talk at the bar after the conference” what his newest book -- Machine Learning Cookbook with Python -- is all about, who it’s for, and the gap it addresses for the community Enjoy the show!   Show Notes: https://ajgoldstein.com/podcast/ep8/ Chris’ Twitter: https://twitter.com/chrisalbon AJ’s Twitter: https://twitter.com/ajgoldstein393/

Our guest this week is flashcard kingpin and former Partially Derivative co-host Chris Albon. Topics of discussion include how good it feels not to have a podcast machine learning flashcards being a "natsec bro" whether Chris would punch a Nazi whether Chris would sexually harass a Nazi whether "magister" is a good woke replacement for "master" whether that Andrew Ng job posting is appropriate and whether any of us would apply for it killing your heroes having a day a week without social media treadmill desks Chris's next podcast and somehow Joel gets going on Harry Potter Please listen to it.

In today's episode, I interview Chris Albon, co-host of Partially Derivative, a fantastic casual discussion podcast about all things data science. Chris is joined by Vidya Spandana and Jonathon Morgan on the show. We discuss the exciting prospects of machine learning and data science in this three part interview! Today's episode is brought to you by Linode. Linode Provides superfast SSD based Linux servers in the cloud starting at $10 a month. Linode is offering Developer Tea listeners $20 worth of credit if you use the code DEVELOPERTEA2017 at checkout. Head over to spec.fm/linode to learn more about what Linode has to offer to Developer Tea listeners .

In today's episode, I interview Chris Albon, co-host of Partially Derivative, a fantastic casual discussion podcast about all things data science. Chris is joined by Vidya Spandana and Jonathon Morgan on the show. We discuss the exciting prospects of machine learning and data science in this three part interview! Today's episode is sponsored by Fuse! Build native iOS and Android apps with less code and better collaboration. Head over to spec.fm/fuse to learn more today!

In today's episode, I interview Chris Albon, co-host of Partially Derivative, a fantastic casual discussion podcast about all things data science. Chris is joined by Vidya Spandana and Jonathon Morgan on the show. We discuss the exciting prospects of machine learning and data science in this three part interview! Today's episode is sponsored by Fuse! Build native iOS and Android apps with less code and better collaboration. Head over to spec.fm/fuse to learn more today!

This month’s podcast features a long journey across Tanzania to discover how low-tech innovation can help developing nations eradicate poverty and meet the Millennium Development Goals. From improved cookstoves to a new system that combines sanitation and composting, small initiatives aimed at local communities can bring big leaps where limited resources and infrastructure are available. Looking at the Millennium Development Goals, we also shed light on what’s needed to reduce global carbon emissions in order to reduce global warming and prompt sustainable development. Joyeeta Gupta of the University of Amsterdam, in the Netherlands, says that countries ignoring the issue of climate change have to be isolated by the international community, while the developing world needs to find a new model of sustainable development. We then talk about conservation with the story of an award-winning vet who is trying to save the endangered grey crowned crane of Rwanda, which is being wiped out by poaching. We also meet Chris Albon of non-profit tech company Ushahidi to discuss how open data tools can map human rights abuses during crises. Finally, there is a sneak preview of next month’s podcast: border water wars and World Water Week.