devopsy.kinsta.cloud

Tag: ml

  • Steady On. We Still Have a Job to Do

    Steady On. We Still Have a Job to Do

    It is all the rage to talk about how significantly large language models (LLMs) and generative AI will change every niche and corner of IT. And change IT it will. But the hype cycle is at its highest right now, and I’d caution against getting too worked up about it. We’re in for a few interesting years, and the spaces I cover (DevOps and security—with a more application-centric perspective but with ops experience) are definitely going to benefit from the various AI bits being worked in. All testing—including AppSec, but all testing—will benefit both in test generation and results analysis/filtering. In fact, I’ll go ahead and say that the question in testing will no longer be, “How much time to we have to devote to it?” and will rapidly become “How many systems’ resources do we have to dedicate to it?” It will still be a long process—AI isn’t speeding up the cumulative time it takes to run thorough testing (be it functional or security). It will make that the bottleneck though, because test generation will at least be LLM-assisted, and will eventually become completely generative AI-created while at the same time, results will get more and more filtering and analysis. A minor result here plus a minor result there will automatically be detected as an extended vulnerability. There are a ton of other possibilities, but these two seem pretty obvious when I look across the testing technology space. More testing and better results analysis are on the horizon—without requiring more staff. Of course, you can do so much more, which would require the organization to add staff to follow up on results, but that’s nothing new; testing finds issues that are interruptive and take man-hours to fix—we all know of bits in the application portfolio that we should probably fix but that aren’t high enough priority to do so, for example. But fixing is different than finding. I have seen demos in security scanners that offer auto-fix options for developers. Run the test during CI, create a ticket and let the developer see suggested code rework right in the IDE. It is early days for this, but if you look, you can find vendors doing it. And that’s really rather cool.

    But we’ve all heard the “this changes everything” storyline over and over. 4GLs, XML, Java and JavaScript were all going to eliminate developers. In fact, they all created more developers. AI has a better chance of actually removing parts of our workload than any of these technologies did, but we’re still in “watch and wait” mode. It seems obvious that grunt coding will fall to LLMs … And aside from keeping staff trained to manage source, that is a good thing. We use open source, modules and libraries to do much the same thing today. I see this as equivalent to entry-level writing jobs. We’ll have to transform from the current model, but the current entry level development model wasn’t ideal anyway, so lets make a better one. Testing capabilities will be greatly enhanced, but testing of all kinds was so far behind new code generation at most orgs that the impact will be all positive as testing catches up. AI-based test generation today can generate thousands of permutations for a given piece of code, and analyze the results to just have analysts look at the truly disconcerting or confusing test results. And the question will become, “How many servers do we want to dedicate to run this?” Which is different, but just means we have more tests running to make a more solid and/or secure product and environment.

    So, in summary, is change coming? Yes. And that is not significantly different than the last couple decades where change has become IT’s constant companion. For most of you, the way you do your job today is not the same as the way you did it even two years ago. ML/GAI/LLMs are just another iteration. The apps still have to be built, deployed and maintained. So until HAL comes along, keep kicking it. You’re keeping it together, and who knows what the future brings.

  • What Are You Securing?

    What Are You Securing?

    Every once in a while, I think it is important to look at the available technologies and tooling and ask, “What are we trying to do?” In the case of information security, we need to ask, “What, exactly, are we trying to protect?”

    The answers to that question are seriously limited. Availability, reputation, data. Since reputation is a business mapping of the other two, that leaves us with availability and data. Of the two, data is more important. Our systems being down absolutely costs us money, but leaking customers’ private data or our corporate secrets, like source code or technology designs, is worse in the long run.

    The problem with these observations is that if we are here to protect the data, and data is the more important piece of security, then why do we have far better application security than data leak prevention? Almost all of you have policies and processes to run an application through the wringer and get an “as secure as possible” result at the end. But data? The purpose of applications is to expose it. And in the process, we risk exposing too much of it or allowing tons of it out of a valid endpoint. I remember in the early days of login-based websites, there was a spectacular leak that let an attacker get hundreds of valid user credentials per page submission (there were a surprising number of these for a couple of years there; this one was at a college). We are facing much the same thing via APIs these days or even via pages meant to make access easier for users.

    Locking all of this behind logins is a useful step, but to use one of the many phrases we’ve stolen from military science, it is only a part of ‘defense in depth.’

    Once it is agreed that data is one of the two key things we need to protect, then the focus should be on how to do so. There are some amazing DLP products, both standalone and integrated into other technologies, that are worth looking into. There are also some great advances in automated response under the guise of AI/ML that should be watched closely because they are changing rapidly. Limiting the maximum number of responses allowed from a query is almost mandatory for application stability and for data leak prevention. And it isn’t that hard to implement.

    I assure you that ne’er-do-wells—particularly state-sponsored and corporate espionage hackers, but also amateurs—are using AI to make attacks more successful and less detectable. The least you can do is respond in kind, with some of the tools that can now do things like notify you about low-and-slow data leak attacks.

    Keep focused on what the goals are. Reiterate them often, and keep rocking it. That’s a lot of data splashing around the organization that your team is responsible for; make sure it’s as safe as possible.

  • AI ‘is Getting Worse’ ¦ AI ‘Will Lose India Jobs’ (Probably Isn’t ¦ Probably Won’t)

    AI ‘is Getting Worse’ ¦ AI ‘Will Lose India Jobs’ (Probably Isn’t ¦ Probably Won’t)

    Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.

    This week, a conundrum: On the one hand, researchers say ChatGPT is losing the plot; and on the other, we hear outsourced coding jobs in India will be replaced by AI. (more…)

  • Applitools Acquires Preflight to Further Apply AI to App Testing

    Applitools Acquires Preflight to Further Apply AI to App Testing

    Applitools this week revealed it has acquired Preflight, which provides a low-code automation platform infused with machine learning algorithms that makes it possible for developers to write their own tests.

    Applitools COO Moshe Milman said Preflight extends the company’s efforts to leverage artificial intelligence (AI) to shift more testing capabilities left toward application developers using its existing test automation platform. The Applitools platform is compatible with frameworks such as Selenium, Playwright and Cypress.

    One of the major challenges organizations encounter as they encourage developers to run more tests is that many of them don’t have the skills required to write them. Applitools’ AI capabilities reduce the cognitive load required to create and run tests, noted Milman.

    The goal isn’t so much to replace dedicated application testing teams but to enable developers to more easily test code as they build it, noted Milman. There will always be a need for dedicated testing teams to evaluate more complex use cases for applications, he added.

    The issue is, thanks to advances in DevOps automation, the pace of application development is outpacing the testing resources available, said Milman. IT teams need to invest in test automation to close what has become a quality gap because not enough code is being thoroughly tested, he noted.

    Test automation also serves to reduce the total cost of creating tests. That enables organizations to conduct more tests without slowing down application development, noted Milman.

    In the longer term, Applitools also plans to leverage large language models (LLMs) to provide generative AI capabilities alongside existing support for machine learning algorithms to further automate testing, noted Milman. Applitools is already making use of computer vision to automate app testing. Preflight adds a set of tools infused with AI to create tests.

    The reasons behind inadequate application testing aren’t entirely clear. Whether it’s because tests are too difficult to create or development teams simply run out of time because of scheduling delays, the impact is the same. Too often, developers convince themselves they will address any issues that might arise with the next update only to become distracted by other priorities. Regardless of the reason, however, there are more poorly-tested applications running in production environments than many DevOps teams would prefer. In fact, much of the technical debt IT organizations need to manage today can be traced back to inadequate testing.

    However, as testing tools become more automated, the overall quality of applications should improve and technical debt may become less of an issue. In the meantime, automated testing promises to improve overall developer productivity as it becomes simpler for developers to create code that has been, at the very least, vetted by algorithms before being added to a codebase. In effect, developers are running out of excuses for not testing applications before delivering them to unsuspecting end users.

  • Must You Collect All That?

    Must You Collect All That?

    It is interesting that we no longer consider teams that super-tool their apps as being slightly ‘off’ or, working in a field where systems were mission critical, even life-and-death. Now, we all are collecting and reporting data points in real-time across not just the SDLC, but across the entire environment. It is wild the amount of data available to the average DevOps team about what is going on with their applications. But it’s not all being used.

    No, seriously, it is unreal. Let me offer two anecdotes that explain why I know you’re not using it all.

    Most of my regular readers know that I worked at F5 for years, the first couple on the DevCentral team, doing the Dev part. Their flagship product, the BIG-IP, collected an incredible amount of super-granular data. I am not exaggerating when I say an incredible amount. They had data points from the wire up to the traffic routes through the load balancing algorithm to nodes. And it was (and likely still is) all available to users via APIs. Whatever data was needed about traffic passing through the BIG-IP or the nodes behind the BIG-IP and the connections in between users and those nodes, you could get.

    One of my projects was to offer a simplified interface to the APIs. They were complex and granular, and customers wanted fast and easy, so I was publishing code that worked as an intermediate layer.

    Note: F5’s APIs, like everyone else’s, are far more advanced today than they were then, and it is easier to do what you need to do because they took the “create a user layer” part seriously. It’s complex machinery so it still has complex APIs, but accessibility is way up from those early days.

    Anyway, BIG-IP collected so much data that I could not name all of the fields, let alone when each one was pertinent. I gave up trying after a while, because if a board-level engineer needed a data point, the idea at the time was “Why not expose it?” so ‘number of connections to a given node’ under this algorithm was hidden among things like ‘microseconds to connect to DNS on that connection’ (I’m making those up; it’s been years). Some were terribly useful to end users and others were not. All were accessed in the same way, and learning by experimenting was the only real way to tell the difference.

    Users did not need all of that data. Users did not need the complexity and obfuscation created by supporting all of that unnecessary data.

    My other example is a massive system of interconnected IoT devices that were centrally connected/controlled with a private, specialized network to manage them (actually, there were several different networks on several different mediums, and that matters where data points are concerned).

    When setting this system up, we estimated about half a million IoT devices with capabilities that ranged from passive monitoring to active controllers of physical devices. Four different communications mechanisms, a couple dozen different devices, every major (and a couple minor) databases and command-and-control software on every major OS, managing from a handful to hundreds of thousands. And all of these were super-tooled because the IoT network was mission-critical. Our databases and the infrastructure to support/protect them were a major part of our data center budget because of the amount of data.

    The amount of data was a flood. The management software was developed by different companies with different transports and endpoints in mind, meaning each had a different dataset. Once all of the networks were running satisfactorily, our next immediate project was to create a unified dataset for reporting and troubleshooting. I was in charge of IT, but the top-level person–because each of those half a million devices was a physical install–was a physical engineer. We fought non-stop about what was important. I was trying to map and convert datapoints between systems; he was trying to maximize data available for reporting up. He often said, “Well no, we don’t need that today, but we might.” This put us in a zone where “number of devices on this system” was on par with “attenuation of the communications line to the nodes on this physical subnet.” Totally unworkable, and our compromise was to put together a small IT team to put a layer over all that raw data and offer an intelligent interface side-by-side with data dumps.

    We are there again today. I hear many of my friends in the enterprise space saying, “Because of ML, we might need that data and storage is cheap.” I disagree. Storage is an overhead that is wasted if you are not actively using the data. We are past the “all data is good data” craze; it’s time to adapt. Useful data that expands insights is all you need tooling for. Report data that will help solve problems, and if the team finds tomorrow that it needs a data point that is not currently collected, it can be added in.

    There is room for debate here–the rate of change for ML systems (please stop calling them AI; they’re not, yet, and we’ll need that word when they are cognizant) means that there is a reasonable argument for “all of the data.” And I’m not proposing limiting data collection. Honestly, application security specifically and enterprise IT security overall has benefitted from datapoint sharing between formerly separate market spaces. If the SCA scan shows X framework is used, then the IAST can narrow down tests to those that look at X and not its competitors, for example. But honestly, don’t drown yourself in a data lake when all you need is a fish tank where you can see all the guppies.

    And keep rocking it. Even if you entered IT two years ago, you are now a seasoned vet and things have changed–beyond all of our expectations, I’d say. Yet you’re still plugging away, keeping the enterprise running. Don’t stop—just try to keep the data hoarding to a reasonable level and make sure the data you have is being used, whether as useful feeds to an ML tool or as consumable information for other people/systems.

  • Standardize: It’s Not the Where. Sometimes it’s Not the What

    Standardize: It’s Not the Where. Sometimes it’s Not the What

    In our industry’s attempts to follow best practices that marketers assure us we must have or we are going to lose to our biggest competitor … No, wait, we’ll lose to that startup no one ever heard of … No, hold on, we’ll lose to foreign competition … No, wait … (okay, I’ll stop. You get the picture), many of us have gotten completely unreasonable in our IT approaches.

    Like asking a vendor, “Where can I run your application?” Well, the more “where” options the better, because every organization has different needs. But a vendor has to start somewhere, and you can tell when they get to a certain size because they don’t want to talk about it. “We’re SaaS,” was the en vogue thing to say six months ago; now we’re back to “containers” or even “cloud,” thanks to a whole slew of cloud-first offerings out there. But what all of these vendors really need to be able to say (and a disappointingly small number of them actually can) is, “We run on containers, so we don’t have to care where you deploy us.”

    Those are the types of scenarios that forward-looking enterprises need to start anticipating—the broadest, easiest-to-implement, longest-lasting solutions that can be found. The days of “devs picked this DB no one ever heard of because it was 0.0000001 microseconds faster on a hundred-line query,” need to be gone. Standardize on highly usable tools across the DevOps toolchain from libraries approved for use in apps to deployment technology.

    A big part of this approach right now is my pretty common refrain: Advanced automation. The more a system can do, the more time staff has to do other important things. All forms of testing are finally within our grasp because of advanced automation technology ranging from Kubernetes programmability and controllers to ‘AI’ filtering of results (Side note: What most IT folks currently call AI is not; we need to stop calling it that and give it a more accurate name. Machine learning is better; advanced automation is more to the point, IMO). But whatever you call it, it enables massive pre-release testing. From functional to security, it offers the ability to make these tests useful by speeding up testing and results processing so staff need only look at issues considered critical. No one cares where the test engine runs, just that it can run tests wherever the application resides.

    The same is true for a whole lot of other ‘whats.’ Vulnerabilities are probably the best example. Tools are starting to look for them in an automated fashion. Very few organizations have a comprehensive list of which vulnerability databases a tool must support. Any given organization might have one or two databases in their list, but most just require that the functionality be available.

    So don’t get bogged down in the details, and don’t select five different tools to do the same job on five different teams. Start standardizing again. Oh, standardization won’t last when something evolutionary comes along in the space, but it will leave you less clutter to clean up and more specific expertise in-house.

    And that leaves you more time to keep wowing the business. Keep those systems running smoothly and keep cranking out more useful stuff for the organization. Most orgs show their appreciation these days, but some still do not. In those cases, we thank you for them. We know what you’re going through—we are going through it, too, and yet the lights are still on. Keep rocking it.

  • How Open Source Can Benefit AI Development

    How Open Source Can Benefit AI Development

    Enterprises are increasingly reliant upon open source software. A full 95% of IT leaders say open source tools are key to their enterprise infrastructure. Simultaneously, we’ve witnessed a sharp increase in the adoption of artificial intelligence (AI) and machine learning (ML). Most IT divisions want to take advantage of AI in some form, and open source technologies will likely be vital to these efforts.

    I recently met with Moses Guttmann, co-founder and CEO of ClearML, to understand the benefits of adopting open source software within the sphere of artificial intelligence. To Guttmann, open source has many benefits—the visibility it grants can equate to helpful feedback from the developer community, and transparency can inform new feature development and identify potential security gaps. There are also countless open source tools to help organizations develop and deploy ML models.

    Below, we’ll consider the benefits and potential drawbacks of incorporating OSS in AI development and explore ways to optimize the use of AI within an organization.

    Benefits of Using Open Source for AI

    AI has evolved in leaps and bounds. What was pure sci-fi a few years ago has become a reality through ChatGPT and other AI-powered platforms. The capabilities of AI are awe-inspiring, but Guttmann believes it still requires fine-tuning to move from entertainment purposes into a valid business driver.

    For example, custom ML models are difficult to construct from scratch, and generic APIs aren’t always that customized to the use case at hand to be valuable. Instead, Guttman foresees more and more organizations downloading open pre-trained models and retraining or customizing them to meet their specific needs.

    For software providers, there are many benefits to adopting an open source strategy. The visibility it creates helps evolve a project with feedback from the community, said Guttmann. This, in turn, helps organizations build better, more stable products. For example, ClearML adopts plenty of OSS to power their user interface, backend and other components, in addition to providing a popular open source CI/CD framework for ML workflows. “We’re true believers in the power of community and like to work with open source and contribute back to these projects,” said Guttman.

    Use Cases for Open Source in AI Development

    Open source software is powering innovation in many ways. And when it comes to AI, there are numerous open source projects to help accelerate AI development. Such projects encompass pre-trained models, data-structuring tools, frameworks to train models and AI deployment pipelines. For example, many open source machine learning frameworks can help engineers build and train their models from scratch. These free-to-use tools include TensorFlow, Keras, PyTorch, H2O, XGBoost and others.

    However, training models from scratch is time-intensive and complicated, said Guttmann. Instead, another option is to use pre-trained models accessible on GitHub or other repositories. Some popular examples are PaddleNLP and Natural Language Toolkit for natural language processing, OpenCV for machine vision and DeepFaceLab for deepfakes. There are also many open source Python packages that can structure and clean data for ML model training purposes.

    Open source software can also be applied to streamline the ML DevOps pipeline itself. AI requires intense processing and is typically deployed to its own computing environments. Infrastructure management around running and training AI is a promising area where open source could assist, said Guttmann. “Automating this process with an abstraction layer to allow data science and ML engineers to access product-grade hardware is a must,” said Guttmann.

    Potential Downsides of Open Source Usage

    Yet, of course, there are some potential downsides to using open source software within the field of AI. Supply chain attacks are still a persistent threat, and numerous open source software risks can present themselves. These include bad actors compromising a legitimate package, name confusion attacks, unmaintained software and untracked dependencies. License restrictions could also potentially block the inclusion of OSS in certain projects.

    Thankfully, the benefits of OSS appear to outweigh these concerns. Automated vulnerability scanning techniques can help detect known issues. It also helps that 89% of technology leaders see enterprise open source as more secure or as secure as proprietary software, according to a Red Hat report.

    Tips For Getting the Most From AI

    Open source is powering digital innovation on many fronts, including AI. But, simply looping in OSS isn’t the only aspect of fully leveraging AI—another will be the operationalization of ML management. The right automation infrastructure in place could enable the agility necessary to continuously retrain models. And since there are potential gaps in the CI/CD process, Guttmann recommended separating AI/ML from your normal CI/CD for security reasons.

    Guttmann also believes that, in general, companies should be more transparent around ML models, including sharing details behind their creation and visibility regarding the datasets fed into them. Overlooking this could lead to ethical concerns. Furthermore, limiting the “hallucinations” generative AI produces is another huge gap that must be bridged, said Guttmann. This could be solved by simply allowing the system to respond with “I don’t know,” instead of confidently spouting nonsense.

    In summary, open source could bring many benefits to bespoke AI/ML adoption. Many mature open source tools exist to enable engineers to create advanced AI and oversee the management and deployment of custom ML models. But the real test will be retaining this open creed by upholding visibility and transparency into the end creations.

  • Looking Ahead, 2023 Edition

    Looking Ahead, 2023 Edition

    DevOps is in a perpetual state of change that keeps all of us engaged and busy, but makes a mess of things overall. Do you cloud-first? Are your APIs GraphQL yet? Are you doing SDN to connect to the VPN so you can put files on the CDN and build closer to the market? Sorry, had to do it. You get the point.

    The point is that the ground is forever changing under us, and often things that seem straightforward are not. Like “cloud-native,” including Kubernetes. Yeah, I get it, you mean software-defined infrastructure that works well with infrastructure-as-code (IaC); this field is confusing enough—calling it cloud is dumb when, these days, cloud has a specific meaning in the minds of IT.

    In our current round of change, a few patterns are showing up. Automation is king. Sometimes it is sold as AI/ML because those are the buzzwords of the day, but it is just continuing the trend of automation to make IT more responsive and predictable. We are rapidly approaching a point where the entirety of the application development/deployment process can be automated reliably and configured to create predictable results. By that I mean that development has more assistance writing code, test has more assistance running tests and processing results, deployment environments can be grabbed from application release automation/orchestration tooling and the final product all of that means that builds can be deployed with very little human interaction unless something breaks.

    Highly automated systems are a prerequisite of high-volume output, though – automobile factories are the easiest example of that statement. My problem is that we don’t seem to be asking “Should we be high-volume producers?” Does the average enterprise require non-stop development? I think that the initial answer is yes, and the more considered answer is “Only because the business wants more, consistently.”

    At some point, the gains we received from DevOps need to be looked at from the “What value is added?” perspective. And that is a business function with IT input. Before SaaS and DevOps made IT easier (to get something running; not minimizing our work here), the business had to consider carefully if it needed that application or that change to an existing one. Today, that consideration is pretty much “Because we want it, put it on the backlog.” But all of our new development adds to our technical debt. While the wave of automation we have been living through since “The Pheonix Project” was first published is keeping us ahead of the curve, we will inevitably need to consider flattening the curve. I suggest that the time is now.

    Getting started with this is actually pretty easy. For many organizations, simply stating, “We do not need two ways to do that” will yield benefits quickly. Other organizations will have to ask “Why can’t we use the functionality of existing tools/apps to do that?” and yet others will need to ask the business, “Do we even need to do that?” Prioritization still exists, and management does an okay job of pushing back at all of the organizations I talk to. Still, at some point it goes beyond “prioritize your wishlist,” and IT managers will have to start asking if a project is even necessary–even if doing so makes waves.

    Continue to rock it, but we’re past time to look for redundancy, synergy and misdirection. There is more complexity coming, and more sweet solutions to problems (both known and unknown) are right around the corner. But, as I’ve said before, simplify as you go because you need to be able to handle the next wave, and eventually, automation won’t cover the gap.

  • 8-Bit Floating Point for AI/ML? | Amazon and Microsoft Shed Tech Jobs

    8-Bit Floating Point for AI/ML? | Amazon and Microsoft Shed Tech Jobs

    Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.

    This week: New ideas bring low-power ML inference, and more big-tech jobs are going. (more…)

  • ClearML Releases New Reports Feature to Share Real-Time Results of Machine Learning Projects and Ignite ML Collaboration Across the Enterprise

    Tel Aviv, IL, January 19, 2023 – ClearML, the leading open source, end-to-end MLOps platform, today announced it has released a new Reports feature that is now generally available. This new feature makes it easy to create and share real-time reports within ClearML as well as connect to third-party editors such as Confluence, Monday, Notion, Colab (Jupyter), and others using embed code. A report can contain charts that auto-update continuously throughout an experiment lifecycle.

    ClearML users can now easily create, collaborate on, and share reports, as well as graphs and charts, in order to summarize and explain experiments and how model versions improve,  analyze results, show experiment comparisons, discuss bugs, demonstrate progress towards milestones, and publish their work.

    “Reporting and showcasing your findings to colleagues, managers, or even your future self is a core component of any modern collaborative workflow,” said Moses Guttmann, Co-founder and CEO of ClearML. “Having one central place where you can seamlessly  report, analyze, plan, and collaborate on your work in real time makes it that much easier. ClearML-Reports connects seamlessly with existing ClearML functionality and is based on markdown, so it’s very easy to export or connect to an external reporting tool.”

    Features and Benefits of ClearML Reports

    The new ClearML Reports feature was designed with a multitude of functionality in mind, including:

    • Creation: It’s now easy to write reports using markdown language, and there is a handy markdown cheatsheet to get started. Reports can be filtered by tags or archived like datasets or experiments.
    • Visualization:  Users can embed scalars, plots, or debug samples from the ClearML experiment manager right into the markdown editor with preview updates. They can even embed descriptive statistics of their datasets or comparisons graphs from multiple experiments — what’s important is that these are not HTML exports, but are still connected to the back-end such that when new data comes in, it will be visible after a report refresh.
    • Team Collaboration: Teammates with ClearML logins can edit reports that are not published as read-only. Reports are also searchable based on their descriptions.
    • Sharing of Reports: Users can share on-platform with others who have a ClearML login and can export to PDF for a time-bound report to share with anyone else. Publishing the report will make it read-only, so it can be used as a single source of truth among colleagues. Users can easily export your reports out of ClearML as markdown or PDF.
    • External Tooling Integration: Many users and their teams each have their own workflows with their own toolchains. By not limiting the reporting functionality to only ClearML reports, users can easily integrate ClearML elements into their favorite third-party reporting tools, like Notion, Confluence, Monday, and others.

    “Reporting our findings on the weekly standups was quite manual,” said Omri Bar, AI Research Team Leader at Theator, the creator of  Surgical Intelligence. “I’d have to jump around between experiments, moving from one plot to another, narrating what was interesting about the graph in question. That’s not very easy for my colleagues to follow along with. Now I can prepare a neat report ahead of time, with all the interesting parts embedded inside, important notes right next to them, cleanly organized by topic. It has made our communication much smoother.”

    See the power of ClearML Reports for yourself by watching the overview video now.

    Get started with ClearML by using our free tier servers or by hosting your own. Read our documentation here. You’ll find more in-depth tutorials about ClearML on our YouTube channel and we also have a very active Slack channel for anyone that needs help. If you need to scale your ML pipelines and data abstraction or need unmatched performance and control, please request a demo. To learn more about ClearML, please visit: https://clear.ml/.

    About ClearML

    ClearML is a unified, open source platform for continuous machine learning (ML), trusted by forward-thinking Data Scientists, ML Engineers, DevOps, and decision makers at leading Fortune 500 companies, enterprises, academia, and innovative start-ups worldwide. We enable customers to build continuous ML workflows — from experiment management and orchestration through data management and scheduling, followed by provisioning and serving — to achieve the fastest time to ML production, fastest time to value, and increased performance. In this way, ClearML accelerates ML adoption across business units, helping companies reach their revenue potential and materialize their ML investments. With thousands of deployments and a vibrant, engaged community, ClearML is transforming the ML space — bridging software, machine learning, and automation. To learn more, visit the company’s website at https://clear.ml.