In corporate finance and investment banking, risk analysis is a crucial job. To assess risk, analysts review research, monitor economic and social conditions, stay informed of regulations, and create models for the investment climate. In short, the inputs into an analysis make for a highly complex and dynamic calculation, one that requires enormous computing power. The vast number of calculations and the way the math is structured typically allows for high degrees of parallelization across many separate processes. To satisfy such a need, grid computing employs any number of machines working together to execute a set of parallelized tasks — which is perfect for risk analysis. By using a networked group of computers that work together as a virtual supercomputer, you can assemble and use vast computer grids for specific time periods and purposes, paying, only for what you use. Also, by splitting tasks over multiple machines, processing time is significantly reduced to increase efficiency and minimize wasted resources.

The Azure Industry Experiences team has recently authored two documents to help those involved in banking scenarios. We show how to implement a risk assessment solution that takes advantage of cloud grid computing technologies.

The first document is a short overview for technical decision makers, especially those considering a burst-to-cloud scenario. The second is a solution guide. It is aimed at solution architects, lead developers and others who want a deeper technical illumination of the strategy and technology.

Recommended next steps

1. Read the Risk Grid Computing Overview
2. Read the Risk Grid Computing Solution Guide

Today, we are happy to introduce support for the Web Authentication specification in Microsoft Edge, enabling better, more secure user experiences and a passwordless experience on the web.

With Web Authentication, Microsoft Edge users can sign in with their face, fingerprint, PIN, or portable FIDO2 devices, leveraging strong public-key credentials instead of passwords.

Staying secure on the web is more important than ever. We trust web sites to process credit card numbers, save addresses and personal information, and even to handle sensitive records like medical information. All this data is protected by an ancient security model—the password. But passwords are difficult to remember, and are fundamentally insecure—often re-used, and vulnerable to phishing and cracking.

For these reasons, Microsoft has been leading the charge towards a world without passwords, with innovations like Windows Hello biometrics and pioneering work with the FIDO Alliance to create an open standard for passwordless authentication – Web Authentication.

We started this journey in 2016, when we shipped the industry’s first preview implementation of the Web Authentication API in Microsoft Edge. Since then, we have been updating our implementation to as we worked with other vendors and the FIDO alliance to develop the standard. In March, the FIDO Alliance announced that the Web Authentication APIs have reached Candidate Recommendation (CR) status in the W3C, a major milestone for the maturity and interoperability of the specification.

Beginning with build 17723, Microsoft Edge supports the CR version of Web Authentication. Our implementation provides the most complete support for Web Authentication to date, with support for a wider variety of authenticators than other browsers.

Windows Hello allows users to authenticate without a password on any Windows 10 device, using biometrics—face and fingerprint recognition—or a PIN number to sign in to web sites. With Windows Hello face recognition, users can log in to sites that support Web Authentication in seconds, with just a glance.

Users can also use external FIDO2 security keys to authenticate with a removable device and your biometrics or PIN. For websites that are not ready to move to a completely passwordless model, backwards compatibility with FIDO U2F devices can provide a strong second factor in addition to a password.

We’re working with industry partners on lighting up the first passwordless experiences around the web. At RSA 2018, we shared a sneak peak of how these APIs could be used to approve a payment on the web with your face. Passwordless authentication experiences like this are the foundation of a world without passwords.

We’re excited to get implementation into the hands of more developers to see what you build. To get started with Web Authentication in Microsoft Edge, check out more information on our implementation in the Web Authentication dev guide, or install Windows Insider Preview build 17723 or higher to try it out for yourself!

– Angelo Liao, Program Manager, Microsoft Edge
– Ibrahim Damlaj, Program Manager, Windows Security

Updated July 30, 2018 10:12 am

Visual Studio Code, or VS Code, is an open source code editor that also includes tools for building and debugging an application. With the Python extension enabled, vscode becomes a great working environment for any Python developer. This article shows you which extensions are useful, and how to configure VS Code to get the most out of it.

If you don’t have it installed, check out our previous article, Using Visual Studio Code on Fedora:

Using Visual Studio Code on Fedora

Install the VS Code Python extension

First, to make VS Code Python friendly, install the Python extension from the marketplace.

Once the Python extension installed, you can now configure the Python extension.

VS Code manages its configuration inside JSON files. Two files are used:

• One for the global settings that applies to all projects
• One for project specific settings

Press Ctrl+, (comma) to open the global settings.

Setup the Python Path

You can configure VS Code to automatically select the best Python interpreter for each of your projects. To do this, configure the python.pythonPath key in the global settings.

// Place your settings in this file to overwrite default and user settings. { "python.pythonPath":"${workspaceRoot}/.venv/bin/python", }

This sets VS Code to use the Python interpreter located in the project root directory under the .venv virtual environment directory.

Use environment variables

By default, VS Code uses environment variables defined in the project root directory in a .env file. This is useful to set environment variables like:

PYTHONWARNINGS="once"

That setting ensures that warnings are displayed when your program is running.

To change this default, set the python.envFile configuration key as follows:

"python.envFile": "${workspaceFolder}/.env",

Code Linting

The Python extension also supports different code linters (pep8, flake8, pylint). To enable your favorite linter, or the one used by the project you’re working on, you need to set a few configuration items.

By default pylint is enabled. But for this example, configure flake8:

"python.linting.pylintEnabled": false, "python.linting.flake8Path": "${workspaceRoot}/.venv/bin/flake8", "python.linting.flake8Enabled": true, "python.linting.flake8Args": ["--max-line-length=90"],

After enabling the linter, your code is underlined to show where it doesn’t meet criteria enforced by the linter. Note that for this example to work, you need to install flake8 in the virtual environment of the project.

Code Formatting

VS Code also lets you configure automatic code formatting. The extension currently supports autopep8, black and yapf. Here’s how to configure black.

"python.formatting.provider": "black", "python.formatting.blackPath": "${workspaceRoot}/.venv/bin/black" "python.formatting.blackArgs": ["--line-length=90"], "editor.formatOnSave": true,

If you don’t want the editor to format your file on save,  set the option to false and use Ctrl+Shift+I to format the current document. Note that for this example to work, you need to install black in the virtual environment of the project.

Running Tasks

Another great feature of VS Code is that it can run tasks. These tasks are also defined in a JSON file saved in the project root directory.

Run a development flask server

In this example, you’ll create a task to run a Flask development server. Create a new Build using the basic template that can run an external command:

Edit the tasks.json file as follows to create a new task that runs the Flask development server:

{ // See https://go.microsoft.com/fwlink/?LinkId=733558 // for the documentation about the tasks.json format "version": "2.0.0", "tasks": [ { "label": "Run Debug Server", "type": "shell", "command": "${workspaceRoot}/.venv/bin/flask run -h 0.0.0.0 -p 5000", "group": { "kind": "build", "isDefault": true } } ] }

The Flask development server uses an environment variable to get the entrypoint of the application. Use the .env file to declare these variables. For example:

FLASK_APP=wsgi.py FLASK_DEBUG=True

Now you can execute the task using Ctrl+Shift+B.

Unit tests

VS Code also has the unit test runners pytest, unittest, and nosetest integrated out of the box. After you enable a test runner, VS Code discovers the unit tests and letsyou to run them individually, by test suite, or simply all the tests.

For example, to enable pytest:

"python.unitTest.pyTestEnabled": true, "python.unitTest.pyTestPath": "${workspaceRoot}/.venv/bin/pytest",

Note that for this example to work, you need to install pytest in the virtual environment of the project.

As global economies move ever closer to a digital future, companies and organizations in every industry vertical are grappling with how to further integrate and deploy technology throughout their business and operations. While Enterprise IT largely led the way, the advantages and lessons learned are now starting to be applied across the board. While the national unemployment rate stands at 4.1%, the overall unemployment rate for tech professionals hit 1.9% in April and the future for open source jobs looks particularly bright. I work in the open source networking space and the innovations and opportunities I’m witnessing are transforming the way the world communicates.

Once a slower moving industry, the networking ecosystem of today — made up of network operators, vendors, systems integrators, and developers — is now embracing open source software and is shifting significantly towards virtualization and software defined networks running on commodity hardware. In fact, nearly 70% of global mobile subscribers are represented by network operator members of LF Networking, an initiative working to harmonize projects that makes up the open networking stack and adjacent technologies.  

Demand for Skills

Developers and sysadmins working in this space are embracing cloud native and DevOps approaches and methods to develop new use cases and tackle the most pressing industry challenges. Focus areas like containers and edge computing are red hot and the demand for developers and sysadmins who can integrate, collaborate, and innovate in this space is exploding.

Open source and Linux makes this all possible, and per the recently published 2018 Open Source Jobs Report, fully 80% of hiring managers are looking for people with Linux skills while 46% are looking to recruit in the networking area and a roughly equal equal percentage cite “Networking” as a technology most affecting their hiring decisions.

Developers are the most sought-after position, with 72% of hiring managers looking for them, followed by DevOps skills (59%), engineers (57%) and sysadmins (49%). The report also measures the incredible growth in demand for containers skills which matches what we’re seeing in the networking space with the creation of cloud native virtual functions (CNFs) and the proliferation of Continuous Integration / Continuous Deployment approaches such as the XCI initiative in the OPNFV.

Get Started

The good news for job seekers in that there are plenty of onramps into open source including the free Introduction to Linux course. Multiple certifications are mandatory for the top jobs so I encourage you to explore the range of training opportunities out there. Specific to networking, check out these new training courses in the OPNFV and ONAP projects, as well as this introduction to open source networking technologies.

If you haven’t done so already, download the 2018 Open Source Jobs Report now for more insights and plot your course through the wide world of open source technology to the exciting career that waits for you on the other side!

Download the complete Open Source Jobs Report now and learn more about Linux certification here.

Chris Bishop leads Microsoft’s research lab in Cambridge, which has been at the forefront of AI, machine learning and deep learning research for 20 years. Its work contributes to many Microsoft products and features, such as the Clutter feature in Office.

Name: Chris Bishop

Role: Technical Fellow and Laboratory Director

Age: 59

Lives: Cambridge, UK

Family: Wife and two sons (both at university, studying Biology and Computer Science)

Pets: Two cats

Hobbies: Flying aeroplanes

Tell us about your current role?

I was one of the first people to join Microsoft’s Research Lab in Cambridge UK, back when the lab was first opened in 1997, before being named Lab Director two-and-a-half years ago, so I’ve been involved in growing and shaping the lab for more than two decades. Today my role includes leadership of the MSR Cambridge lab, as well as coordination of the broader Microsoft presence in Cambridge. I am fortunate in being supported by a very talented leadership team and a highly capable and motivated team of support staff.

What were your previous jobs?

My background is in theoretical physics. After graduating from Oxford, I did a PhD in quantum field theory at the University of Edinburgh, exploring some of the fundamental mathematics of matter, energy, and space-time. After my PhD I wanted to do something that would have potential for practical impact, so I joined the UK’s national fusion research lab to work on the theory of magnetically confined plasmas as part of a long-term goal to create unlimited clean energy. It was during this time that there were some breakthroughs in the field of neural networks. I was very inspired by the concept of machine intelligence, and the idea that computers could learn for themselves. Initially I started applying neural networks to problems in fusion research, and we became the first lab to use neural networks for real-time feedback control of a high-temperature fusion plasma.

In fact, I found neural networks so fascinating that, after about eight years working on fusion research, I took a rather radical step and switched fields into machine learning. I became a Professor at Aston University in Birmingham, where I set up a very successful research lab. Then I took a sabbatical and came to Cambridge for six months to run a major, international programme called “Neural Networks and Machine Learning” at the Isaac Newton Institute. The programme started on July 1, 1997, on the very same day that Microsoft announced it was opening a research lab in Cambridge, its first outside the US. I was approached by Microsoft to join the new lab, and have never looked back.

What are your aims at Microsoft?

My ambition is for the lab to have an impact on the real world at scale by tackling very hard research problems, and by leveraging the advantages and opportunities we have as part of Microsoft. I often say that I want the MSR Cambridge lab to be a critical asset for the company.

I’m also very passionate about diversity and inclusion, and we have introduced multiple initiatives over the last year to support this. We are seeing a lot of success in bringing more women into technical roles in the lab, across both engineering and research, and that’s very exciting to see.

What’s the hardest part of your job?

A core part of my job is to exercise judgment in situations where there is no clear right answer. For instance, in allocating limited resources I need to look at the risk, the level of investment, the potential for impact, and the timescale. At any one time there will be some things we are investing in that are quite long term but where the impact could be revolutionary, along with other things that have perhaps been researched for several years which are beginning to get real traction, all the way to things that have had real-world impact already. The hardest part of my job is to weigh up all these factors and make some difficult decisions on where to place our bets.



What’s the best part of your job?

The thing I enjoy most is the wonderful combination of technology and people. Those are two aspects I find equally fascinating, yet they offer totally different kinds of challenges. We, as a lab, are constantly thinking about technology, trends and opportunities, but also about the people, teams, leadership, staff development and recruitment, particularly in what has become a very competitive talent environment. The way these things come together is fascinating. There is never a dull day here.

What is a leader?

I think of leadership as facilitating and enabling, rather than directing. One of the things I give a lot of attention to is leadership development. We have a leadership team for the lab and we meet once a week for a couple of hours. I think about the activities of that team, but also about how we function together. It’s the diversity of the opinions of the team members that creates a value that’s greater than the sum of its parts. Leadership is about harnessing the capabilities of every person in the lab and allowing everyone to bring their best game to the table. I therefore see my role primarily as drawing out the best in others and empowering them to be successful.

What are you most proud of?

Last year I was elected a Fellow of the Royal Society, and that was an incredibly proud moment. There’s a famous book I got to sign, and you can flip back and see the signatures of Isaac Newton, Charles Darwin, Albert Einstein, and pretty much every scientist you’ve ever heard of. At the start of the book is the signature of King Charles II who granted the royal charter, so this book contains over three-and-a-half centuries of scientific history. That was a very humbling but thrilling moment.

Another thing I’m very proud of was the opportunity to give the Royal Institution Christmas Lectures. The Royal Institution was set up more than 200 years ago – Michael Faraday was one of the early directors – and around 14 Nobel prizes have been associated with the Institution, so there is a tremendous history there too. These days it’s most famous for the Christmas Lectures, which were started by Faraday. Ever since the 1960s these lectures have been broadcast on national television at Christmas, and I watched them as a child with my mum and dad. They were very inspirational for me and were one of the factors that led me to choose a career in science. About 10 years ago, I had the opportunity to give the lectures, which would have been inconceivable to me as a child. It was an extraordinary moment to walk into that famous iconic theatre, where Faraday lectured many times and where so many important scientific discoveries were first announced.

One Microsoft anecdote that relates to the lectures was that getting selected was quite a competitive process. It eventually came down to a shortlist of five people, and I was very keen to be chosen, especially as it was the first time in the 200 year history of the lectures that they were going to be on the subject of computer science. I was thinking about what I could do to get selected, so I wrote to Bill Gates, explained how important these lectures were and asked him whether, if I was selected, he would agree to join me as a guest in one of the lectures. Fortunately, he said yes, and so I was able to include this is my proposal to the Royal Institution. When I was ultimately selected, I held Bill to this promise, and interviewed him via satellite on live television during one of the lectures.

What inspires you?

I love the idea that through our intellectual drive and curiosity we can use technology to make the world a better place for millions of people. For example, the field of healthcare today largely takes a one-size-fits-all approach that reactively waits until patients become sick before responding, and which is increasingly associated with escalating costs that are becoming unsustainable. The power of digital technology offers the opportunity to create a data-driven approach to healthcare that is personalised, predictive and preventative, and which could significantly reduce costs while also improving health and wellbeing. I’ve made Healthcare AI one of the focal points of the Cambridge lab, and I find it inspiring that the combination of machine learning, together with Microsoft’s cloud, could help to bring about a much-needed transformation in healthcare.

What is your favourite Microsoft product?

A few years ago, the machine learning team here in Cambridge built a feature, in collaboration with the Exchange team, called Clutter. It sorts out the email you should pay attention to now, from the ones that can be left to, say, a Friday afternoon. I love it because it’s used by tens of millions of people, and it has some very beautiful research ideas at the heart of it – something called a hierarchical Bayesian machine learning model. This gives it a nice out-of-the-box experience, a sort of average that does OK for everybody, but as you engage with it, it personalises and learns your particular preferences of what constitutes urgent versus non-urgent email. The other reason I’m particularly fond of it is that when I became Lab Director, the volume of email in my inbox quadrupled. That occurred just as we were releasing the Clutter feature, so it arrived just in time to save me from being overwhelmed.

What was the first bit of technology that you were excited about?

When I was a child I was very excited about the Apollo moon landings. I was at an age where I could watch them live on television and knew enough to understand what an incredible achievement they were. Just think of that Saturn launch vehicle that’s 36 storeys high, weighs 3,000 tonnes, is burning 15 tonnes of fuel a second, and yet it’s unstable. So, it must be balanced, rather like balancing a broom on your finger, by pivoting those massive engines backwards and forwards on hydraulic rams in response to signals from gyroscopes at the top of the rocket. It’s that combination of extreme brute force with exquisite precision, along with dozens of other extraordinary yet critical innovations, that made the whole adventure just breath-taking. And the filtering algorithms used by the guidance system are an elegant application of Bayesian inference, so it turns out that machine learning is, literally, rocket science.

Tags: Apollo, cambridge, Chris Bishop, Fellow, microsoft, research