• PyData July: CNNs & NLP

    HolidayCheck Group AG

    18:30 Doors Open & Pizza 19:00 Welcome to PyData 19:15 CNNs by Kamaljeet Singh 20:00 Introduction to NLP by Thejeswi Preetham CNNs by Kamaljeet Singh Talk Abstract: Convolutional Neural Networks (CNNs) are the state of the art in image recognition benchmarks (such as ImageNet) and have successfully been applied to numerous vision tasks ranging from robotics, autonomous driving to health care or games (AlphaGo). This talk will mainly focus on a related use-case of CNNs that leverages the benefits of transfer learning to achieve highly accurate results in the domain of assessing the quality of images. We will look at Google's NIMA (Neural Image Assessment) architecture together with its implementation in two popular Deep Learning libraries, Keras and fastai, the latter of which gains more and more traction in the community due to its integration of powerful learning enhancement methods and convenient Data Block API. Finally, we will have a brief look at more recent approaches for semi-supervised learning where the same or better performance can be achieved with far less labelled training data. Speaker Bio: Kamaljeet Singh is a Data Scientist at HolidayCheck focusing on applying Deep Learning methods in the travel domain. He completed his Masters degree in Computer Science with a specialization in Cognitive Technical Systems at the University of Freiburg in 2018. He considers himself to be a Deep Learning fanboy and extremely fascinated at how the recent developments in AI have brought major breakthroughs in many technological areas. Introduction to NLP by Thejeswi Preetham Talk Abstract: From the onset of computers we have always wanted to communicate in natural language (eg. English) with it. This has been achieved to some extent with voice assistants, chatbots and semantic text searches. Applications can do this using natural language processing by analyzing and processing large amounts of language data. This talk will cover some fundamental methods for extracting meaning out of plain text and tools to do so. Speaker Bio: Thejeswi Preetham is a Python programmer and data analyst. He has a lot of experience with NLP, automation, and web apps. Best wishes, Thanks to HolidayCheck (https://www.holidaycheck.de/) for sponsoring the food & drinks!

    5
  • Massive Parallel and Elastic AI Pipelines with TensorFlow and Kubernetes

    TNG Technology Consulting GmbH

    17:30 Welcome and Pizza 18:00 Fast Track Intro to DeepLearning 18:30 What's new in TensorFlow[masked]:15 Build Enterprise AI Pipelines using KubeFlow Sponsored by: 1) The IBM Data Science Community https://www.ibm.com/community/datascience, where data scientists and developers to learn, share, and engage with their peers and industry renowned data scientists => Pizza 2) IBM Codait (Center for Open Source Data and AI Technologies) www.codait.org => Speaker Travel 3) TNG Technology Consulting https://www.tngtech.com => Drinks, Location Talk Summary: Toward the end of 2015, Google released TensorFlow, which started out as just another numerical library, but has grown to become a de facto standard in AI technologies. TensorFlow received a lot of hype as part of its initial release, in no small part because it was released by Google. Despite the hype, there have been complaints on usability. Especially, for example, the fact that debugging was only possible after construction of the static execution graph. In addition to that, neural networks needed to be expressed as a set of linear algebra operations which was considered too low level by many practitioners. PyTorch and Keras addressed many of the flaws in TensorFlow and gained a lot of ground. TensorFlow 2.0 successfully addressed those complaints and promises to become the go-to framework for many AI problems. Romeo Kienzler introduces you to the most prominent changes in TensorFlow 2.0 and how you can use these new features successfully in your projects. He explores eager execution, parallelization strategies, the advantages of the tight high-level Keras integration, live neural network training monitoring using TensorBoard, automated hyper parameter optimization, model serving with TensorFlow service, TensorFlow.js, and TensorFlow Lite. He shares an outlook on TFX—where Google is planning to open source its complete AI pipeline—and contrasts it with existing de facto standard frameworks like Apache Spark. Speaker Bio: Romeo Kienzler is Chief Data Scientist at the IBM Center for Open Source Data and AI Technologies (CODAIT) in San Fransisco, owning the strategy lead for AI Model Training. He works as Associate Professor for artificial intelligence at the Swiss University of Applied Sciences Berne and his current research focus is on cloud-scale machine learning and deep learning using open source technologies including TensorFlow, Keras, DeepLearning4J, Apache SystemML and the Apache Spark stack. He also contributes to various open source projects. He regularly speaks at international conferences including significant publications in the area of data mining, machine learning and Blockchain technologies. Location TNG Beta-Straße 13a 85774 Unterföhring 17:30 Welcome and Pizza 18:00 What's new in TensorFlow[masked]:30 Build Enterprise AI Pipelines using KubeFlow

    4
  • PyData June: Talks on NetworkX & Apache Airflow

    E.ON Energy Projects GmbH

    PyData Munich is back with a couple of accessible data science tutorials! Schedule: - 18:30 Doors open - 19:00 Welcome to PyData Munich - 19:10 “Introduction to Graph Data with NetworkX” by Roland Rodde @ E.On In this talk, Roland introduces graph data, showing how to use the NetworkX Python package to generate, load, and measure graphs using various graph metrics in order to help people better intuit the general feeling behind various graph metrics and feel more comfortable working with graphs. - 19:55 “Workflow Management Tools & Apache Airflow” by Martin Wurzer, Data Engineer @ KI labs In this talk, Martin tries to address the following questions: - What are workflow management systems? - Why should you use them to organise your data pipelines? - How can you use Apache Airflow to schedule, organise and monitor your workflows? Martin works as a Data Engineer at KI labs. In his free time, he likes to build cool side projects. - 20:40 Job advertisements and networking with beer and pizza Many thanks to E.On (https://www.eon.de/de/pk.html) for hosting and sponsoring this event! Hope to see you there! Nick & Nithish P.S.: If you are interested in giving a talk or hosting an event, please contact one of the organizers! If you are looking for inspiration for a talk topic or want to request one, please check out the ever-growing lists of topics here: https://docs.google.com/document/d/1wrdv27B7UGRJelbk8EQdHmBZ8uuO0I4q2JlJrsiCGlI/edit?usp=sharing

  • PyData April: Talks on Kotlin & Computer Vision

    JetBrains Event Space

    PyData Munich is back with a couple of accessible data science tutorials! Schedule: - 18:30 Doors open - 19:00 Welcome to PyData Munich - 19:10 “Kotlin 💕 Data Science?” by Preslav Rachev, Senior Backend Engineer @ KI labs Software engineering and data science are two very different disciplines. Accommodating both in the same project is not trivial. The key to success relies on developers and data scientists having a middle ground where they could do their work productively, and without affecting the other side. Preslav's talk will try to revisit Kotlin as a language potentially suiting both sides. He will reflect on its advantages, and also the current roadblocks. Preslav is a Java/Python double agent with over 10 years of cumulative programming experience (both in industry and academia). He loves open-source software, making music, writing, and distance running. - 19:55 “Introduction to Computer Vision” by Anton Kasyanov, Senior Machine Learning Engineer @ DataRobot The talk is aimed at the engineers who are curious about the area and have no prior experience with it. The main goal is to give the feel of what is it all about, what problems can be solved and some basic methods to do that. Anton works as a Senior Machine Learning Engineer in DataRobot for almost two years. Prior to that, he did a lot of backend and ML development using Python. He got his Computer Science Masters degree in RWTH Aachen (Germany) with a focus on Computer Vision. - 20:40 Job advertisements and networking with beer and pizza Many thanks to JetBrains (https://www.jetbrains.com/) for hosting and sponsoring this event! Hope to see you there! Connor & Nithish P.S.: If you are interested in giving a talk or hosting an event, please contact one of the organizers! If you are looking for inspiration for a talk topic or want to request one, please check out the ever-growing lists of topics here: https://docs.google.com/document/d/1wrdv27B7UGRJelbk8EQdHmBZ8uuO0I4q2JlJrsiCGlI/edit?usp=sharing

    1
  • Python Talk about PyTorch & Error Bars in Data Vizualisations

    PyData Munich is back with another accessible data science tutorial at another new venue, Data Reply Talk 1: "Understanding PyTorch: PyTorch in image processing" by Dmitrii Azarnykh Description: The talk will demonstrate the key PyTorch functionality and show how it can make our lives easier, especially with a focus on statistical analysis and image processing. We will start our journey from the introduction to automatic differentiation: what this animal is and how it is used in backpropagation. Then we will move to the basics of PyTorch: how to initialize tensors, to compute gradients, and we will also build a linear regression model with PyTorch together. At the last part of the talk, we will look at how to classify images of broken and valid objects using built-in PyTorch functionality. If for some of you, the terminology mentioned sounds unfamiliar, you’re in a right place to start understanding it, as we will start from basics. Talk 2: "Error Bars are Your Experiment: the Statistics Behind the Most Important Feature of Data Visualizations" by Nick Del Grosso Description: Your choice of error bar type has a big influence on the message you send to your audience when presenting data. In this talk, we'll cover the fundamental intuitions behind the standard deviation, standard error of the mean, and confidence intervals, and apply that intuition to both the data visualization approaches used by Seaborn and the fundamental concepts behind Bayesian modeling in PyMC3. Schedule: - 18:30 Doors open - 19:00 Welcome to PyData Munich - 19:10 Understanding PyTorch: PyTorch in image processing by Dmitrii Azarnykh - 19:55 "Error Bars are Your Experiment: the Statistics Behind the Most Important Feature of Data Visualizations" - 20:30 Job advertisements and networking with beer and pizza In case, anyone is interested in giving a lightning talk or sharing something you like, get in touch with us. Many thanks to Data Reply (http://www.reply.com) for hosting and sponsoring this event! Hope to see you there! Nithish & Connor P.S.: If you are interested in giving a talk or hosting an event, please contact one of the organizers! If you are looking for inspiration for a talk topic or want to request one, please check out the ever-growing lists of topics here: https://docs.google.com/document/d/1wrdv27B7UGRJelbk8EQdHmBZ8uuO0I4q2JlJrsiCGlI/edit?usp=sharing

    3
  • PyData February

    Motius GmbH

    PyData Munich is back with another accessible data science tutorial! Talk to be announced! Schedule: - 18:30 Doors open - 19:00 Welcome to PyData Munich - 19:15 “Introduction to Micropython” by Sebastian Plamauer, Embedded Engineer & Tech Specialist @ Motius - 19.40 “Introduction to Apache Pig – Data Engineering for Python Programmers” by Furqan Shakoor, Software Engineer & Tech Specialist @ Motius - 20.05 "How to track and organize your experimentation process" by Jakub Czakon You will learn how with some simple steps you can have your work organized around creative iterations, reproducible and easy to share with anyone. You will see how to easily track the code, metrics, hyperparameters, learning curves, data versions and more. Bonus point: we will speak with a bot that knows a lot about the experiments. - 20:30 Job advertisements and networking with beer and pizza Many thanks to Motius (https://www.motius.de/) for hosting and sponsoring this event! Hope to see you there! Connor P.S.: If you are interested in giving a talk or hosting an event, please contact one of the organizers! If you are looking for inspiration for a talk topic or want to request one, please check out the ever-growing lists of topics here: https://docs.google.com/document/d/1wrdv27B7UGRJelbk8EQdHmBZ8uuO0I4q2JlJrsiCGlI/edit?usp=sharing

    4
  • Python Talk about Snakemake

    codecentric AG

    PyData Munich is back with another accessible data science tutorial in the New Year at a new space, Codecentric Snakemake in the "Wild", by Joshua Friedman - Why should I use a workflow engine? - What is Snakemake? - Why should I use Snakemake? - What can be done with Snakemake? Join this scintillating talk to answer these questions and more! Joshua is a Data Engineer turned Data Product Manager at KI labs. His background is extremely atypical spanning many fields - Civil Engineering to Coastal Engineering to Remote Sensing to Big Data to ... He has exactly 6.82 years of experience ranging from scientific research to deep learning to consulting. He likes hacking on his pi, automating his life with shortcuts and putting QR codes everywhere. Schedule: - 18:30 Doors open - 19:00 Welcome to PyData Munich - 19:15 Main Talk on SnakeMake - 20:05 Job advertisements and networking with beer and pizza In case, anyone is interested in giving a lightning talk or sharing something you learned during the holidays, get in touch with us. Many thanks to Codecentric (https://www.codecentric.de/) for hosting and sponsoring this event! Hope to see you there! Nithish & Connor P.S.: If you are interested in giving a talk or hosting an event, please contact one of the organizers! If you are looking for inspiration for a talk topic or want to request one, please check out the ever-growing lists of topics here: https://docs.google.com/document/d/1wrdv27B7UGRJelbk8EQdHmBZ8uuO0I4q2JlJrsiCGlI/edit?usp=sharing

    5
  • Python Talks about Building a Blockchain and REST API

    JetBrains Event Space

    PyData Munich is back with another couple accessible data science tutorials, once again at the JetBrains Event Space! There will be three talks: Talk 1: A lightening talk by Ernst Haagsman, PyCharm PMM - JetBrains about some brand-new data science-oriented functionality in PyCharm. Talk 2: Let's build a Python Blockchain, by Riswan Lathif Description: The attendees would be able to understand what a Blockchain is, what it is not, and why there is so much hype surrounding it after the talk. A small prototype blockchain would be built over the course of the session with Python. No prior experience with Blockchain technology is required. Riswan is a Masters Informatics student at TUM and a backend engineer at KI Labs. He has worked with SAP Labs and DLR in the past. He is currently doing a thesis on DAG based blockchain alternatives. He likes tweaking open source projects and attending hackathons. Talk 3: "REST API" by Victor Movileanu Description: The main concepts of RESTful web services will be highlighted, as well as how to use them based on the bitcoin trading website Kraken. At the end of the talk, a trading bot for learning purposes could be created. Victor Movileanu studied Biochemistry and learned programming by himself during his masters, which he used a lot during his internships and master thesis for data analysis. Currently, he is interested in web development (Django framework, Angular 6). Schedule: - 18:30 Doors open - 19:00 Welcome to PyData Munich - 19:05 A lightening talk by JetBrains: some brand-new data science-oriented functionality in PyCharm - 19:15 Talk Riswan Lathif: Blockchain basics and hands on - 20:10 Talk Victor Movileanu: REST API and bitcoin trading - 21:05 Job advertisements and networking with beer and pizza Many thanks to JetBrains (https://www.jetbrains.com//) for hosting and sponsoring this event! Hope to see you there! Riswan, Jane & Maria P.S.: If you are interested in giving a talk or hosting an event, please contact one of the organizers! If you are looking for inspiration for a talk topic or want to request one, please check out the ever-growing lists of topics here: https://docs.google.com/document/d/1wrdv27B7UGRJelbk8EQdHmBZ8uuO0I4q2JlJrsiCGlI/edit?usp=sharing

    2
  • Python Talks About Bayesian Inference

    Burda Bootcamp

    PyData Munich is back after the Summer break with another couple of accessible data science tutorials, once again at the Burda Bootcamp There will be two talks: - Talk 1: Speaker: Connor Leahy Title: "Things I Wish I'd Been Told When I First Started Learning Programming" Description: To start off the new year of PyData talks, Connor Leahy will bring newcomers and veterans together by getting us talking about the journey to learn programming! Using humourous anecdotes from the programming and data science worlds, he'll point out the pitfalls, and share pro tips for where to look for useful information! - Talk 2: Speaker: Mohammad Bashiri Title: "Bayesian Inference with Python" Description: What is a probability distribution and how do I use it to model data? In this tutorial, Mohammad Bashiri will teach us how we can intuit the process of inferring information from data by recognizing the components of Bayes' theorem, and how it can be applied to fit data and generate predictions using the Python package PyMC3! Schedule: * 18:30 Doors open * 19:00 Welcome to PyData Munich * 19:05 Talk Connor Leahy: Things I Wish I'd Been Told When I First Started Learning Programming * 19:45 Talk Mohammad Bashiri: Bayesian Inference with Python * 20:30 Job Advertisements and Networking with beer Many thanks to Burda Bootcamp (http://burdabootcamp.de) for hosting and sponsoring this event! Hope to see you there! Connor, Nick & Nithish P.S.: If you are interested in giving a talk or hosting an event, please contact one of the organizers! If you are looking for inspiration for a talk topic or want to request one, please check out the ever-growing lists of topics here: https://docs.google.com/document/d/1wrdv27B7UGRJelbk8EQdHmBZ8uuO0I4q2JlJrsiCGlI/edit?usp=sharing

    10
  • Python Talks About Multidimensional Data Analysis

    JetBrains Event Space

    PyData Munich is back with another couple accessible data science tutorials, once again at the JetBrains Event Space! There will be three talks: - Talk: “Machine Learning with Python: Clustering with Scikit-Learn” by Parastou Kohvaei: We'll be revisiting one of Python's most powerful machine learning packages, Scikit-Learn. During this introductory-level tutorial, we'll learn how to use this package to perform data clustering and pattern recognition. - Talk: “Multidimensional Data Analysis” by Joseph Donovan: Data often exists in a multi-dimensional space, and when the number of dimensions is high visualization and analysis can prove challenging. This talk will demonstrate approaches for handling, visualizing, and reducing the dimensionality of such datasets. We’ll also briefly cover some useful parts of sklearn and xarray (think N-dimensional pandas dataframes). - Talk: "Understanding Numpy's 'as.strided': Advanced array manipulation" by Nick Del Grosso: Numpy's dtype, shape, and stride parameters work to produce multidimensional arrays. In this talk, Nick will demonstrate how these parameters can be used as a versatile tool for high-speed, copy-less data analysis, especially when using Numpy's stride_tricks subpackage. Schedule (Attention: we start a bit earlier!): * 18:15 Doors open * 18:40 Welcome to PyData Munich * 18:45 Talk Parastou Kohvaei: Machine Learning with Python: Clustering with Scikit-Learn * 19:25 Talk Joseph Donovan: Multidimensional Data Analysis * 20:05 Talk Nick Del Grosso: Understanding Numpy's 'as.strided': Advanced array manipulation * 20:30 Job Advertisements and Networking with pizza & beer Many thanks to JetBrains (https://www.jetbrains.com//) for hosting and sponsoring this event! Hope to see you there! This is our last event of the first season. We are going to a summer break and will be back with new strength for the next season of awesome Python talks around October. Maria & Nick P.S.: If you are interested in giving a talk or hosting an event, please contact one of the organizers (including myself: [masked] or Nick: [masked])! If you are looking for inspiration for a talk topic or want to request one, please check out the ever-growing lists of topics here: https://docs.google.com/document/d/1wrdv27B7UGRJelbk8EQdHmBZ8uuO0I4q2JlJrsiCGlI/edit?usp=sharing

    7