The Python Podcast.__init__

The majority of machine learning projects that you read about or work on are built around batch processes. The model is trained, and then validated, and then deployed, with each step being a discrete and isolated task. Unfortunately, the real world is rarely static, leading to concept drift and model failures. River is a framework for building streaming machine learning projects that can constantly adapt to new information. In this episode Max Halford explains how the project works, why you might (or might not) want to consider streaming ML, and how to get started building with River.

Deep learning is a revolutionary category of machine learning that accelerates our ability to build powerful inference models. Along with that power comes a great deal of complexity in determining what neural architectures are best suited to a given task, engineering features, scaling computation, etc. Predibase is building on the successes of the Ludwig framework for declarative deep learning and Horovod for horizontally distributing model training. In this episode CTO and co-founder of Predibase, Travis Addair, explains how they are reducing the burden of model development even further with their managed service for declarative and low-code ML and how they are integrating with the growing ecosystem of solutions for the full ML lifecycle.

Machine learning has the potential to transform industries and revolutionize business capabilities, but only if the models are reliable and robust. Because of the fundamental probabilistic nature of machine learning techniques it can be challenging to test and validate the generated models. The team at Deepchecks understands the widespread need to easily and repeatably check and verify the outputs of machine learning models and the complexity involved in making it a reality. In this episode Shir Chorev and Philip Tannor explain how they are addressing the problem with their open source deepchecks library and how you can start using it today to build trust in your machine learning applications.

Building an ML model is getting easier than ever, but it is still a challenge to get that model in front of the people that you built it for. Baseten is a platform that helps you quickly generate a full stack application powered by your model. You can easily create a web interface and APIs powered by the model you created, or a pre-trained model from their library. In this episode Tuhin Srivastava, co-founder of Basten, explains how the platform empowers data scientists and ML engineers to get their work in production without having to negotiate for help from their application development colleagues.

Starting a new project is always exciting and full of possibility, until you have to set up all of the repetitive boilerplate. Fortunately there are useful project templates that eliminate that drudgery. PyScaffold goes above and beyond simple template repositories, and gives you a toolkit for different application types that are packed with best practices to make your life easier. In this episode Florian Wilhelm shares the story behind PyScaffold, how the templates are designed to reduce friction when getting a new project off the ground, and how you can extend it to suit your needs. Stop wasting time with boring boilerplate and get straight to the fun part with PyScaffold!

Application configuration is a deceptively complex problem. Everyone who is building a project that gets used more than once will end up needing to add configuration to control aspects of behavior or manage connections to other systems and services. At first glance it seems simple, but can quickly become unwieldy. Bruno Rocha created Dynaconf in an effort to provide a simple interface with powerful capabilities for managing settings across environments with a set of strong opinions. In this episode he shares the story behind the project, how its design allows for adapting to various applications, and how you can start using it today for your own projects.

Software is eating the world, but that code has to have hardware to execute the instructions. Most people, and many software engineers, don't have a proper understanding of how that hardware functions. Charles Petzold wrote the book "Code: The Hidden Language of Computer Hardware and Software" to make this a less opaque subject. In this episode he discusses what motivated him to revise that work in the second edition and the additional details that he packed in to explore the functioning of the CPU.

The generation, distribution, and consumption of energy is one of the most critical pieces of infrastructure for the modern world. With the rise of renewable energy there is an accompanying need for systems that can respond in real-time to the availability and demand for electricity. FlexMeasures is an open source energy management system that is designed to integrate a variety of inputs intelligently allocate energy resources to reduce waste in your home or grid. In this episode Nicolas Höning explains how the project is implemented, how it is being used in his startup Seita, and how you can try it out for your own energy needs.

How And Why To Build Effective Teams As An Engineering Leader

Working on hardware projects often has significant friction involved when compared to pure software. Brian Pugh enjoys tinkering with microcontrollers, but his "weekend projects" often took longer than a weekend to complete, so he created Belay. In this episode he explains how Belay simplifies the interactions involved in developing for MicroPython boards and how you can use it to speed up your own experimentation.

Static typing versus dynamic typing is one of the oldest debates in software development. In recent years a number of dynamic languages have worked toward a middle ground by adding support for type hints. Python's type annotations have given rise to an ecosystem of tools that use that type information to validate the correctness of programs and help identify potential bugs. At Instagram they created the Pyre project with a focus on speed to allow for scaling to huge Python projects. In this episode Shannon Zhu discusses how it is implemented, how to use it in your development process, and how it compares to other type checkers in the Python ecosystem.

Every software project is subject to a series of decisions and tradeoffs. One of the first decisions to make is which programming language to use. For companies where their product is software, this is a decision that can have significant impact on their overall success. In this episode Sean Knapp discusses the languages that his team at Ascend use for building a service that powers complex and business critical data workflows. He also explains his motivation to standardize on Python for all layers of their system to improve developer productivity.

Writing code is only one piece of creating good software. Code reviews are an important step in the process of building applications that are maintainable and sustainable. In this episode On Freund shares his thoughts on the myriad purposes that code reviews serve, as well as exploring some of the patterns and anti-patterns that grow up around a seemingly simple process.

Quality assurance in the software industry has become a shared responsibility in most organizations. Given the rapid pace of development and delivery it can be challenging to ensure that your application is still working the way it's supposed to with each release. In this episode Jonathon Wright discusses the role of quality assurance in modern software teams and how automation can help.

The goal of every software team is to get their code into production without breaking anything. This requires establishing a repeatable process that doesn't introduce unnecessary roadblocks and friction. In this episode Ronak Rahman discusses the challenges that development teams encounter when trying to build and maintain velocity in their work, the role that access to infrastructure plays in that process, and how to build automation and guardrails for everyone to take part in the delivery process.

Every startup begins with an idea, but that won't get you very far without testing the feasibility of that idea. A common practice is to build a Minimum Viable Product (MVP) that addresses the problem that you are trying to solve and working with early customers as they engage with that MVP. In this episode Tony Pavlovych shares his thoughts on Python's strengths when building and launching that MVP and some of the potential pitfalls that businesses can run into on that path.

Application architectures have been in a constant state of evolution as new infrastructure capabilities are introduced. Virtualization, cloud, containers, mobile, and now web assembly have each introduced new options for how to build and deploy software. Recognizing the transformative potential of web assembly, Matt Butcher and his team at Fermyon are investing in tooling and services to improve the developer experience. In this episode he explains the opportunity that web assembly offers to all language communities, what they are building to power lightweight server-side microservices, and how Python developers can get started building and contributing to this nascent ecosystem.

As your code scales beyond a trivial level of complexity and sophistication it becomes difficult or impossible to know everything that it is doing. The flow of logic and data through your software and which parts are taking the most time are impossible to understand without help from your tools. VizTracer is the tool that you will turn to when you need to know all of the execution paths that are being exercised and which of those paths are the most expensive. In this episode Tian Gao explains why he created VizTracer and how you can use it to gain a deeper familiarity with the code that you are responsible for maintaining.

Analysis of streaming data in real time has long been the domain of big data frameworks, predominantly written in Java. In order to take advantage of those capabilities from Python requires using client libraries that suffer from impedance mis-matches that make the work harder than necessary. Bytewax is a new open source platform for writing stream processing applications in pure Python that don't have to be translated into foreign idioms. In this episode Bytewax founder Zander Matheson explains how the system works and how to get started with it today.

Building a fully functional web application has been growing in complexity along with the growing popularity of javascript UI frameworks such as React, Vue, Angular, etc. Users have grown to expect interactive experiences with dynamic page updates, which leads to duplicated business logic and complex API contracts between the server-side application and the Javascript front-end. To reduce the friction involved in writing and maintaining a full application Sam Willis created Tetra, a framework built on top of Django that embeds the Javascript logic into the Python context where it is used. In this episode he explains his design goals for the project, how it has helped him build applications more rapidly, and how you can start using it to build your own projects today.

Virtually everything that you interact with on a daily basis and many other things that make modern life possible were designed and modeled in software called CAD or Computer-Aided Design. These programs are advanced suites with graphical editing environments tailored to domain experts in areas such as mechanical engineering, electrical engineering, architecture, etc. While the UI-driven workflow is more accessible, it isn't scalable which opens the door to code-driven workflows. In this episode Jeremy Wright discusses the design, uses, and benefits of the CadQuery framework for building 3D CAD models entirely in Python.

Building any software project is going to require relying on dependencies that you and your team didn't write or maintain, and many of those will have dependencies of their own. This has led to a wide variety of potential and actual issues ranging from developer ergonomics to application security. In order to provide a higher degree of confidence in the optimal combinations of direct and transitive dependencies a team at Red Hat started Project Thoth. In this episode Fridolín Pokorný explains how the Thoth resolver uses multiple signals to find the best combination of dependency versions to ensure compatibility and avoid known security issues.

Most developers have encountered code completion systems and rely on them as part of their daily work. They allow you to stay in the flow of programming, but have you ever stopped to think about how they work? In this episode Meredydd Luff takes us behind the scenes to dig into the mechanics of code completion engines and how you can customize them to fit your particular use case.

Russell Keith-Magee is an accomplished engineer and a fixture of the Python community. His work on the Beeware suite of projects is one of the most ambitious undertakings in the ecosystem and unfailingly forward-looking. With his recent transition to working for Anaconda he is now able to dedicate his full focus to the effort. In this episode he reflects on the journey that he has taken so far, how Beeware is helping to address some of the threats to Python's long term viability, and how he envisions its future in light of the recent release of PyScript, an in-browser runtime for Python.

Digital cameras and the widespread availability of smartphones has allowed us all to generate massive libraries of personal photographs. Unfortunately, now we are all left to our own devices of how to manage them. While cloud services such as iPhotos and Google Photos are convenient, they aren't always affordable and they put your pictures under the control of large companies with their own agendas. LibrePhotos is an open source and self-hosted alternative to these services that puts you in control of your digital memories. In this episode the maintainer of LibrePhotos, Niaz Faridani-Rad, explains how he got involved with the project, the capabilities that it offers for managing your image library, and how to get your own instance set up to take back control of your pictures.

Investing effectively is largely a game of information access and analysis. This can involve a substantial amount of research and time spent on finding, validating, and acquiring different information sources. In order to reduce the barrier to entry and provide a powerful framework for amateur and professional investors alike Didier Rodrigues Lopes created the OpenBB Terminal. In this episode he explains how a pandemic project that started as an experiment has led to him founding a new company and dedicating his time to growing and improving the project and its community.

The experimentation phase of building a machine learning model requires a lot of trial and error. One of the limiting factors of how many experiments you can try is the length of time required to train the model which can be on the order of days or weeks. To reduce the time required to test different iterations Rolando Garcia Sanchez created FLOR which is a library that automatically checkpoints training epochs and instruments your code so that you can bypass early training cycles when you want to explore a different path in your algorithm. In this episode he explains how the tool works to speed up your experimentation phase and how to get started with it.

Programmers love to automate tedious processes, including refactoring your code. In order to support the creation of code modifications for your Python projects Jimmy Lai created LibCST. It provides a richly typed and high level API for creating and manipulating concrete syntax trees of your source code. In this episode Jimmy Lai and Zsolt Dollenstein explain how it works, some of the linting and automatic code modification utilities that you can build with it and how to get started with using it to maintain your own Python projects.

Communication is a fundamental requirement for any program or application. As the friction involved in deploying code has gone down, the motivation for architecting your system as microservices goes up. This shifts the communication patterns in your software from function calls to network calls. In this episode Idit Levine explains how the Gloo platform that she and her team at Solo have created makes it easier for you to configure and monitor the network topologies for your microservice environments. She also discusses what developers need to know about networking in cloud native environments and how a combination of API gateways and service mesh technologies allow you to more rapidly iterate on your systems.

Cloud native architectures have been gaining prominence for the past few years due to the rising popularity of Kubernetes. This introduces new complications to development workflows due to the need to integrate with multiple services as you build new components for your production systems. In order to reduce the friction involved in developing applications for cloud native environments Michael Schilonka created Gefyra. In this episode he explains how it connects your local machine to a running Kubernetes environment so that you can rapidly iterate on your software in the context of the whole system. He also shares how the Django Hurricane plugin lets your applications work closely with the Kubernetes process model.

Science is founded on the collection and analysis of data. For disciplines that rely on data about the earth the ability to simulate and generate that data has been growing faster than the tools for analysis of that data can keep up with. In order to help scale that capacity for everyone working in geosciences the Pangeo project compiled a reference stack that combines powerful tools into an out-of-the-box solution for researchers to be productive in short order. In this episode Ryan Abernathy and Joe Hamman explain what the Pangeo project really is, how they have integrated a combination of XArray, Dask, and Jupyter to power these analytical workflows, and how it has helped to accelerate research on multidimensional geospatial datasets.

A common piece of advice when starting anything new is to "begin with the end in mind". In order to help the engineers at Wayfair manage the complete lifecycle of their applications Joshua Woodward runs a team that provides tooling and assistance along every step of the journey. In this episode he shares some of the lessons and tactics that they have developed while assisting other engineering teams with starting, deploying, and sunsetting projects. This is an interesting look at the inner workings of large organizations and how they invest in the scaffolding that supports their myriad efforts.

Kubernetes is a framework that aims to simplify the work of running applications in production, but it forces you to adopt new patterns for debugging and resolving issues in your systems. Robusta is aimed at making that a more pleasant experience for developers and operators through pre-built automations, easy debugging, and a simple means of creating your own event-based workflows to find, fix, and alert on errors in production. In this episode Natan Yellin explains how the project got started, how it is architected and tested, and how you can start using it today to keep your Python projects running reliably.

Building a machine learning application is inherently complex. Once it becomes necessary to scale the operation or training of the model, or introduce online re-training the process becomes even more challenging. In order to reduce the operational burden of AI developers Robert Nishihara helped to create the Ray framework that handles the distributed computing aspects of machine learning operations. To support the ongoing development and simplify adoption of Ray he co-founded Anyscale. In this episode he re-joins the show to share how the project, its community, and the ecosystem around it have grown and evolved over the intervening two years. He also explains how the techniques and adoption of machine learning have influenced the direction of the project.

As software projects grow and change it can become difficult to keep track of all of the logical flows. By visualizing the interconnections of function definitions, classes, and their invocations you can speed up the time to comprehension for newcomers to a project, or help yourself remember what you worked on last month. In this episode Scott Rogowski shares his work on Code2Flow as a way to generate a call graph of your programs. He explains how it got started, how it works, and how you can start using it to understand your Python, Ruby, and PHP projects.

One of the most persistent challenges faced by organizations of all sizes is the recording and distribution of institutional knowledge. In technical teams this is exacerbated by the need to incorporate technical review feedback and manage access to data before publishing. When faced with this problem as an early data scientist at AirBnB, Chetan Sharma helped create the Knowledge Repo project as a solution. In this episode he shares the story behind its creation and growth, how and why it was released as open source, and the features that make it a compelling option for your own team's knowledge management journey.

Software development is a complex undertaking due to the number of options available and choices to be made in every stage of the lifecycle. In order to make it more scaleable it is necessary to establish common practices and patterns and introduce strong opinions. One area that can have a huge impact on the productivity of the engineers engaged with a project is the tooling used for building, validating, and deploying changes introduced to the software. In this episode maintainers of the Pants build tool Eric Arellano, Stu Hood, and Andreas Stenius discuss the recent updates that add support for more languages, efforts made to simplify its adoption, and the growth of the community that uses it. They also explore how using Pants as the single entry point for all of your routine tasks allows you to spend your time on the decisions that matter.

It doesn't matter how amazing your application is if you are unable to deliver it to your users. Frustrated with the rampant complexity involved in building and deploying software Vlad A. Ionescu created the Earthly tool to reduce the toil involved in creating repeatable software builds. In this episode he explains the complexities that are inherent to building software projects and how he designed the syntax and structure of Earthly to make it easy to adopt for developers across all language environments. By adopting Earthly you can use the same techniques for building on your laptop and in your CI/CD pipelines.

The process of getting software delivered to an environment where users can interact with it requires many steps along the way. In some cases the journey can require a large number of interdependent workflows that need to be orchestrated across technical and organizational boundaries, making it difficult to know what the current status is. Faced with such a complex delivery workflow the engineers at Ericsson created a message based protocol and accompanying tooling to let the various actors in the process provide information about the events that happened across the different stages. In this episode Daniel Ståhl and Magnus Bäck explain how the Eiffel protocol allows you to build a tooling agnostic visibility layer for your software delivery process, letting you answer all of your questions about what is happening between writing a line of code and your users executing it.

When we are creating applications we spend a significant amount of effort on optimizing the experience of our end users to ensure that they are able to complete the tasks that the system is intended for. A similar effort that we should all consider is optimizing the developer experience for ourselves and other engineers who contribute to the projects that we work on. Adam Johnson recently wrote a book on how to improve the developer experience for Django projects and in this episode he shares some of the insights that he has gained through that project and his work with clients to help you improve the experience that you and your team have when collaborating on software development.

Pandas has grown to be a ubiquitous tool for working with data at every stage. It has become so well known that many people learn Python solely for the purpose of using Pandas. With all of this activity and the long history of the project it can be easy to find misleading or outdated information about how to use it. In this episode Matt Harrison shares his work on the book "Effective Pandas" and some of the best practices and potential pitfalls that you should know for applying Pandas in your own work.

Developers hate wasting effort on manual processes when we can write code to do it instead. Cog is a tool to manage the work of automating the creation of text inside another file by executing arbitrary Python code. In this episode Ned Batchelder shares the story of why he created Cog in the first place, some of the interesting ways that he uses it in his daily work, and the unique challenges of maintaining a project with a small audience and a well defined scope.

Statistical regression models are a staple of predictive forecasts in a wide range of applications. In this episode Matthew Rudd explains the various types of regression models, when to use them, and his work on the book "Regression: A Friendly Guide" to help programmers add regression techniques to their toolbox.

Every software project needs a tool for managing the repetitive tasks that are involved in building, running, and deploying the code. Frustrated with the limitations of tools like Make, Scons, and others Eduardo Schettino created doit to handle task automation in his own work and released it as open source. In this episode he shares the story behind the project, how it is implemented under the hood, and how you can start using it in your own projects to save you time and effort.

Whether we like it or not, advertising is a common and effective way to make money on the internet. In order to support the work being done at Read The Docs they decided to include advertisements on the documentation sites they were hosting, but they didn't want to alienate their users or collect unnecessary information. In this episode David Fischer explains how they built the Ethical Ads network to solve their problem, the technical and business challenges that are involved, and the open source application that they built to power their network.

Podcasts are one of the few mediums in the internet era that are still distributed through an open ecosystem. This has a number of benefits, but it also brings the challenge of making it difficult to find the content that you are looking for. Frustrated by the inability to pick and choose single episodes across various shows for his listening Wenbin Fang started the Listen Notes project to fulfill his own needs. He ended up turning that project into his full time business which has grown into the most full featured podcast search engine on the market. In this episode he explains how he build the Listen Notes application using Python and Django, his work to turn it into a sustainable business, and the various ways that you can build other applications and experiences on top of his API.

Making Orbital Mechanics More Accessible With Poliastro

Deep learning frameworks encourage you to focus on the structure of your model ahead of the data that you are working with. Ludwig is a tool that uses a data oriented approach to building and training deep learning models so that you can experiment faster based on the information that you actually have, rather than spending all of our time manipulating features to make them match your inputs. In this episode Travis Addair explains how Ludwig is designed to improve the adoption of deep learning for more companies and a wider range of users. He also explains how the Horovod framework plugs in easily to allow for scaling your training workflow from your laptop out to a massive cluster of servers and GPUs. The combination of these tools allows for a declarative workflow that starts off easy but gives you full control over the end result.

A lot of time and energy goes into data analysis and machine learning projects to address various goals. Most of the effort is focused on the technical aspects and validating the results, but how much time do you spend on considering the experience of the people who are using the outputs of these projects? In this episode Benn Stancil explores the impact that our technical focus has on the perceived value of our work, and how taking the time to consider what the desired experience will be can lead us to approach our work more holistically and increase the satisfaction of everyone involved.

The true power of artificial intelligence is its ability to work collaboratively with humans. Nate Joens co-founded Structurely to create a conversational AI platform that augments human sales teams to help guide potential customers through the initial steps of the funnel. In this episode he discusses the technical and social considerations that need to be combined for a seamless conversational experience and how he and his team are tackling the problem.

Every machine learning model has to start with feature engineering. This is the process of combining input variables into a more meaningful signal for the problem that you are trying to solve. Many times this process can lead to duplicating code from previous projects, or introducing technical debt in the form of poorly maintained feature pipelines. In order to make the practice more manageable Soledad Galli created the feature-engine library. In this episode she explains how it has helped her and others build reusable transformations that can be applied in a composable manner with your scikit-learn projects. She also discusses the importance of understanding the data that you are working with and the domain in which your model will be used to ensure that you are selecting the right features.