Feed aggregator

From the iPhone 16 and Google Pixel 9 to the Samsung Galaxy S25 series, here are the best 5G phones that we've tested and think you should buy.

It's Memorial Day. To honor the fallen military service members this year, the Up First newsletter asked readers to share stories of their loved ones.

(Image credit: Win McNamee/Getty Images)

Quickly find the last time you called your mom on your iPhone -- and reminder: You should probably call your mom.

• More Google Pixel 10 leaks emerge
• Colors and wallpapers have appeared online
• The phones should launch in August

We've already seen the Google Pixel 10 being filmed for an advert on the streets of Canada, but the leaks aren't stopping: we now have unofficial information about some of the color options and wallpapers the flagship phone will bring with it.

According to tipster Mystic Leaks (via 9to5Google), the standard Pixel 10 will be available in Obsidian (black), Blue, Iris (purple), and Limoncello (yellow-ish) shades. Limoncello could be similar to the Lemongrass option we saw with the Google Pixel 7 in 2022.

As for the Pixel 10 Pro and the Pixel 10 Pro XL, the colors listed here are Obsidian (black), Green, Sterling (Gray), and Porcelain (white-ish). We've only got the names though, and there are no images showing what these colors actually look like.

The Pixel 9 offers Obsidian (black), Peony (pink), Wintergreen (green-ish), and Porcelain (white-ish). The Pro and Pro XL models come in Obsidian (black), Porcelain (white-ish), Rose Quartz (pink), and Hazel (gray).

Pixel 10 wallpapers

The Google Pixel 9 Pro XL (Image credit: Peter Hoffmann)

From the same source, we've got a host of high-resolution Pixel 10 wallpapers, and the team at Android Authority has collected them all together in a bundle, so you can install them on your current phone if you'd like to.

There are a lot of swirls and shapes and gradients here, and everything is very abstract. The colors of the backdrops also match the leaked colors of the phones, and each image has both a dark and a light option to match Android's visual modes.

We haven't heard too much about the Google Pixel 10 so far, apart from what was spotted at the recent promotional shoot, but it is expected to show up sometime in August – perhaps with a display upgrade and a significant speed boost.

Before then, Android 16 should begin to make its way out to the masses. The software is bringing with it numerous improvements, and is going to introduce a significant visual overhaul known as Material 3 Expressive.

You might also like

• How the Pixel 10 can beat the iPhone
• The best Pixels you can buy right now
• Rumored upgrades for the Pixel 10

• Marks & Spencer suffered a cyber-incident in April 2025
• Reports claimed the attack was the work of ScatteredSpider
• Tata Consultancy Services is investigating if the attack came from its network

Tata Consultancy Services (TCS), an Indian IT company and part of the massive Tata Group conglomerate, is currently investigating whether the recent cyberattack on Marks & Spencer (M&S) originated from its infrastructure.

In late April 2025, M&S confirmed suffering a “cyber incident” which affected its stores and resulted in changes to store operations.

Later reports said the company had to take some of its systems and processes offline, and was forced to disable contactless and Click and Collect services in stores, since the incident was, in fact, a ransomware attack. Online orders were also halted. The disruption persisted for weeks, M&S’ market capitalization dropped by £1 billion, and customer data was allegedly stolen by the actors.

Get Keeper Personal for just $1.67/month, Keeper Family for just $3.54/month, and Keeper Business for just $7/month

​Keeper is a cybersecurity platform primarily known for its password manager and digital vault, designed to help individuals, families, and businesses securely store and manage passwords, sensitive files, and other private data.

It uses zero-knowledge encryption and offers features like two-factor authentication, dark web monitoring, secure file storage, and breach alerts to protect against cyber threats.

Preferred partner (What does this mean?)View Deal

Targeting Tata

It had been reported the group known as Scattered Spider was behind the ordeal

Now, BBC News reports TCS, which has been servicing M&S for more than a decade, is investigating whether it was the stepping stone to the attack. Right now, both parties are staying silent, but the investigation should wrap up before June 2025.

TCS is part of the large Indian conglomerate Tata Group, which counts more than 100 companies across a wide range of industries. As such, it is a major target for all sorts of cybercriminals, and roughly two years ago, Hive Ransomware struck Tata Power, India’s largest integrated power company. Early this year, Tata Technologies, a global engineering services provider was also attacked.

The attack is reportedly the work of Scattered Spider, a ransomware organization usually targeting UK retailers, financial institutions, technology firms, and entertainment/gambling organizations. The group is not as tightly-knit as organizations such as LockBit or Cl0p.

It is relatively loose, and operates within a larger hacking community known as “the Com”. Its members engage in all kinds of attacks, from social engineering and SIM swapping, to ransomware.

We have reached out to TCS for comment and will update the article if we hear back.

Via BBC

You might also like

• Marks & Spencer outage allegedly linked to ScatteredSpider ransomware attack
• Take a look at our guide to the best authenticator app
• We've rounded up the best password managers

In today's digital economy, the ability to handle explosive growth without performance ramifications isn't just a technical consideration, it's now a business imperative. So when success arrives, systems must be ready.

Throughout my career advising technology and business leaders, I've witnessed a recurring scenario: a company experiences unexpected success - perhaps they had a successful viral marketing campaign, or suddenly face market interest, or even a random rapid uptick in customer adoption - only to have this triumph transform into a technical crisis as systems falter under the load and stress.

What should be a celebratory moment instead becomes an emergency. Performance levels considerably dip. Customer experience suffers. And the very success that should propel the business forward becomes its biggest operational challenge.

This phenomenon isn't limited to startups and it isn’t necessarily new. Established enterprises frequently encounter these issues during product launches, seasonal peaks, or when entering new markets. Black Friday becomes a nightmare for fresh retailers. The root cause is rarely insufficient hardware or lack of technical talent. More often than not it is that the architectural foundations weren't designed for rapid, unpredictable scaling.

Why traditional approaches can fail

Conventional technology stacks typically perform well under predictable, linear growth conditions. However, real-world business expansion is rarely so simple. Life and business comes in surges, spikes, and sometimes happens overnight.

Traditional databases particularly struggle with these dynamics. When transaction volumes multiply, these systems often hit performance bottlenecks that can't be resolved by simply adding more hardware and their scalability is limited by the biggest available box. Connection limits are reached, query performance deteriorates, and the costs for digital infrastructure climb without delivering proportional benefits.

This is a particular headache that many players in the cryptocurrency space face, where market volatility can trigger 5x transaction volume increases within a matter of minutes. Platforms built on rigid architectures simply cannot adapt quickly enough, leading to trading halts or poor functionality precisely when users need reliability most.

Similarly problematic are monolithic architectures, which are geared for initial speed-to-market rather than long-term flexibility. These approaches might launch quickly, but they rarely support sustainable hypergrowth.

Building from the ground up

Forward-thinking companies are increasingly adopting architectures specifically designed for unpredictable scaling patterns. At the core of this approach is the need for horizontal scalability. This is the ability to expand capacity by adding instances rather than continuously upgrading to larger, more expensive IT infrastructure. In short, flexibility and adaptability is prioritized.

One cryptocurrency exchange that we've worked with demonstrates this principle effectively. By implementing a distributed database architecture, they maintain sub-millisecond response times even during market volatility. So if a run on a certain coin dramatically leads to substantial trading volume fluctuations and customer demand, their platform can automatically scale to handle this without any impacts to the overall service offering.

Equally important is the adoption of cloud-native design patterns - be it deployed on a public cloud, private cloud or just on premises. Microservices, containerization, and orchestration tools all allow businesses to scale cloud computing components independently and recover quickly from failures or setbacks. This modularity essentially supports innovation without compromising stability.

Data model flexibility also plays a crucial role. When another trading platform needed to quickly add new cryptocurrencies to their exchange, their flexible schema approach allowed them to introduce new assets without database migrations or downtime. Understandably, this is a critical advantage in the fast-moving digital asset space.

What does this mean for technology leaders?

For executives preparing their organizations for potential hypergrowth, four priorities consistently make the difference. Firstly, they must design for horizontal scaling from day one. Systems should be built to scale outwardly, but not only in an upward direction. This approach will provide long-term resilience and cost efficiency, something that becomes increasingly valuable as businesses grow and develop.

Secondly, leaders should look to embrace automation. The past two decades have shown how manual processes rarely manage to scale well. Investing in automated provisioning, deployment, and monitoring will not only reduce errors, it allows engineering talent to focus on innovation rather than firefighting issues.

On top of this, they have to stress test beyond their expected peaks. Many systems fail because they're only tested to their current limits. Rigorous testing at 5-10x anticipated peak loads helps identify bottlenecks before they have the chance to impact customers.

Lastly, every leader seeking out hypergrowth must outline architectural efficiency as a key boardroom focus. What I mean here is that scaling isn't purely around performance. It’s as much about financial sustainability, meaning everything from granular resource management to efficient data architecture can help maintain steady growth.

The competitive edge of scalability

In markets where digital experience defines success, scalability is no longer just a technical consideration, it has to be a strategic business capability.

The most successful organizations recognize this. Technology foundations either enable or constrain their ability to capitalize on opportunities and markets. Thinking of recent conversations I’ve had with cryptocurrency executives, when markets surge with interest, exchanges with truly scalable architectures will be the ones that welcome new customers seamlessly. Competitors not following this approach will be forced to implement emergency registration freezes or they risk crumbling altogether.

Ultimately, scaling isn't just about handling growth. It's about being prepared for success, whenever and however that arrives. The question isn't whether your business will face a scaling challenge, but whether you'll be ready when opportunity presents itself.

We've compiled a list of the best cloud databases.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

Under the Biden administration, the Consumer Financial Protection Bureau finalized a rule barring medical debt from appearing on credit reports. Now, the agency is siding with the credit industry groups suing to have the rule vacated.

(Image credit: Saul Loeb)

President Trump has ordered a Veterans Affairs campus in West Los Angeles to house 6,000 homeless vets by 2028, but details are elusive.

(Image credit: Quil Lawrence)

In 2016, Tulika Prasad was at the grocery store checkout line with her seven-year-old son, who is non-verbal and autistic. A woman understood what was going on when Prasad's son had an outburst.

(Image credit: Ravish Kumar)

A performance of the masterpiece will be transmitted into space on Saturday. The waltz has been associated with space travel since its inclusion in the film 2001: A Space Odyssey.

(Image credit: ESA)

Science is an economic driver in Hamilton, Mont., thanks to Rocky Mountain Laboratories, a federal research lab. Now, layoffs and funding cuts are having an impact in this town far from Washington.

(Image credit: Katheryn Houghton/KFF Health News)

Food apps can help you figure out what's in your food and whether it's nutritious. Just scan the barcode on the packet with your phone. But different apps can give very different results. Here's why.

(Image credit: AJ_Watt)

As the proud owner of an iPhone 15 for almost two years, I've had no issues with the handset since I bought it. It runs perfectly for my needs – music, YouTube, texting and aimless doomscrolling social media – and seamlessly integrates with my other Apple devices.

Other than its Pro siblings and a handful of Android competitors, the iPhone 15 was top of the line when I bought it. I’d just been paid, so I plonked down AU$1,499 ($799 / £799) to purchase it outright to replace my broken iPhone 12 Mini.

(That’s a purchase I cringe at after experiencing the value on offer from the best cheap phones, but I digress…)

The iPhone 16 marked a larger upgrade over its predecessor than usual thanks to the addition of Apple Intelligence – even if its launch has been less than smooth, with many parts of the promised Siri upgrade still up in the air.

Still, the iPhone 15 is an excellent smartphone in 2025, which is why it caught my eye when I found it for AU$1,077 here in Australia where I'm based (which converts to around $692 / £519). There's similarly enticing deals abroad, too – like in the US, where it's just $100 when switching to T-Mobile.

(Image credit: Zachariah Kelly / TechRadar)

However, in February, Apple threw a curveball at the iPhone 15 when it introduced another option for Cupertino loyalists looking to save on an upgrade when it launched the iPhone 16e for $599 / £599 / AU$999.

Like the iPhone 5c and the three iPhone SE models, this new ‘budget’ Apple handset has made small concessions to keep the price down, while still allowing buyers to purchase a truly new iPhone that can access the latest iOS features. The iPhone 16e is arguably even more enticing than its SE forebears, as it offers the power to handle Apple Intelligence.

This creates an interesting conundrum – if I needed a new phone and didn't want to splurge on the iPhone 16, which device is the better choice: the iPhone 15, or iPhone 16e?

TechRadar has an entire iPhone 16e vs iPhone 15 comparison article based on this question and it concludes that, for most people, it's worth spending a little extra and go for the older iPhone 15.

But after spending a week with the iPhone 16e, I disagree.

Better battery, baby

(Image credit: Future / Lance Ulanoff)

For many smartphone buyers, camera quality is key – but for me, battery life is far more important and the iPhone 16e dominates the iPhone 15 in this category.

While Apple doesn't disclose exact battery capacity, third-party reporting shows that the iPhone 16e has a 3,961mAh battery compared to the iPhone 15's 3,349mAh.

It's not just that larger size that makes the 16e longer lasting either. The iPhone 16e's C1 cellular chip – which is exclusive to the device – processes power more efficiently, resulting in a significantly improved stamina.

This was very noticeable in my time with it. Granted, my iPhone 15's battery capacity is slightly degraded down to 91% these days, but I limit its overnight charging to stop at 85%. As a result, after about three hours of listening, watching, scrolling and texting, my iPhone 15's often sitting at less than 30% by 9:30am.

It's 3:30pm as I write this, and with the same battery settings and general screen-on time, the iPhone 16e I'm currently using is sitting at 44%.

My experience seems to fully back up Apple’s own claims, with the brand boasting that the iPhone 16e offers 26 hours of video playback – 15% better than the iPhone 16's 22 hours and a 23% increase over the iPhone 15.

Apple Intelligence is already pretty smart, actually

Having fun creating AI-generated images in Apple's Playground app (Image credit: Future/Jacob Krol)

We're still waiting for AI Siri – and Apple might have to let users swap Siri for another default voice assistant party alternatives – I was surprised by how much I enjoyed Apple Intelligence on the iPhone 16e, a set of features the iPhone 15 lacks.

Visual Intelligence is helpful, letting you quickly search for or ask ChatGPT about any object you take a photo of. And the Clean Up feature is useful for removing photo bombers or objects from any given image, like Samsung's similar Object Remover tool as found on newest Galaxy devices.

And, while I rarely used them, I appreciated the (mostly) constructive AI-generated message replies and smarter phrasing suggestions. Highlighting your written text opens an array of AI-powered options by clicking the Apple Intelligence logo (or 'Writing Tools'). In any app it can proofread or rewrite your text to sound more friendly, professional or concise.

Image 1 of 3

Visual Intelligence analyzing potato chips (Image credit: Future)Image 2 of 3

Using Apple Intelligence to create a Genmoji of a dragon holding a hot dog (Image credit: Future)Image 3 of 3

Using AI to editing and proofread text messages (Image credit: Future / Max Delaney)

Moreover, and especially helpful when writing up notes, is its ability to format text into key points, a list or a table. You also have the option to compose text with ChatGPT.

However, I think my favorite thing about Apple Intelligence is the ability to create my own emojis. Called Genmojis, it lets you turn anything – like my own face and other regularly found faces in my camera roll, or a highland cow surrounded by flowers – into an emoji or sticker.

As someone who uses emojis quite sparingly, I'm now a Genmoji-making dynamo. While the AI tools and features of the iPhone 16 family are far from revolutionary, they're both fun and generally useful. It's a small but significant advantage for the iPhone 16e over the iPhone 15.

Bring the action

(Image credit: Future / Lance Ulanoff)

The last little feature that I think puts the 16e above the 15 is the Action Button. It, like Apple Intelligence, is exclusive to the iPhone 15 Pro, Pro Max and the iPhone 16 series.

This handy little button replaces the mute/silent switch from older iPhones. There's nothing revolutionary here: all it does is offer shortcuts for commonly used features like Silent Mode, Focus, Camera, Visual Intelligence, Torch and any other app, like Instagram.

Personally, I didn't find myself using any of those preset options, and instead set the Action Button to control my Do Not Disturb mode.

It's such a small difference – after all, unlocking the device, bringing up the Control Center and activating Focus is hardly a laborious task. However, it's a small quality-of-life change that I thoroughly appreciated – letting me turn it on without even directly looking at my phone.

Winner by a split decision

The two phones are nearly identical apart from the camera array (Image credit: Future / Max Delaney)

The iPhone 16e vs iPhone 15 contest is by no means a knockout by the newer model. There are two main reasons that the older iPhone may be the better choice for some people: display and camera.

The iPhone 16e only has a single 48MP Fusion camera, while the iPhone 15 pairs a 48MP main camera with a 12MP ultrawide lens that's equally useful for grand nature shots and trying to fit the whole family into one photo. More importantly, the 16e's single lens means you can't take silly up-close photos of your friends or dog with the 0.5x zoom.

The 15 also has a (small) lead on the 16e in terms of display, as the latter reverts back to the iPhone 14’s notched display rather than the Dynamic Island found on subsequent devices. Personally, I don't mind it, but for some users it could be the reason to spend a little more for the iPhone 15. The latter’s display is brighter and (slightly) higher res too – 1179 x 2556 with a max brightness of 2,000 nits compared to the 16e's 1170 x 2532 and 1,200 nits.

MagSafe charging is also missing from the iPhone 16e. It was rumored this was to make room for the C1 chip, but that has since been denied by Apple according to Macworld. The 16e can still wirelessly charge, but it lacks the magnet.

I'd never much required MagSafe until I recently purchased a magnetic power bank – which is now all but useless with the iPhone 16e. And users who have a magnetic car mount will probably sorely miss this functionality.

The iPhone 15 still has a place, then, and it's a wonderful purchase if you can get it for close to the same price as the iPhone 16e.

It's still ultimately more expensive than its new sibling, though – and unless you really need a telephoto lens, I think the iPhone 16e is the budget iPhone to have.

You might also like...

• The best iPhone 2025: which Apple smartphone reigns supreme?
• The iPhone Fold is apparently being tested with a feature we’ve been waiting years for
• iPhone 17 Air leak hints at how the super-slim phone might solve its potential battery life issues

As businesses realized the potential of artificial intelligence (AI), the race began to incorporate machine learning operations (MLOps) into their commercial strategies. But integrating machine learning (ML) into the real world proved challenging, and the vast gap between development and deployment was made clear. In fact, research from Gartner tells us 85% of AI and ML fail to reach production.

In this piece, we’ll discuss the importance of blending DevOps best practices with MLOps, bridging the gap between traditional software development and ML to enhance an enterprise’s competitive edge and improve decision-making with data-driven insights. We’ll expose the challenges of separate DevOps and MLOps pipelines and outline a case for integration.

Challenges of Separate Pipelines

Traditionally, DevOps and MLOps teams operate with separate workflows, tools, and objectives. Unfortunately, this trend of maintaining distinct DevOps and MLOps pipelines leads to numerous inefficiencies and redundancies that negatively impact software delivery.

1. Inefficiencies in Workflow Integration

DevOps pipelines are designed to optimize the software development lifecycle (SDLC), focusing on continuous integration, continuous delivery (CI/CD), and operational reliability.

While there are certainly overlaps between the traditional SDLC and that of model development, MLOps pipelines involve unique stages like data preprocessing, model training, experimentation, and deployment, which require specialized tools and workflows. This distinct separation creates bottlenecks when integrating ML models into traditional software applications.

For example, data scientists may work on Jupyter notebooks, while software engineers use CI/CD tools like Jenkins or GitLab CI. Integrating ML models into the overall application often requires a manual and error-prone process, as models need to be converted, validated, and deployed in a manner that fits within the existing DevOps framework.

2. Redundancies in Tooling and Resources

DevOps and MLOps have similar automation, versioning, and deployment goals, but they rely on separate tools and processes. DevOps commonly leverages tools such as Docker, Kubernetes, and Terraform, while MLOps may use ML-specific tools like MLflow, Kubeflow, and TensorFlow Serving.

This lack of unified tooling means teams often duplicate efforts to achieve the same outcomes.

For instance, versioning in DevOps is typically done using source control systems like Git, while MLOps may use additional versioning for datasets and models. This redundancy leads to unnecessary overhead in terms of infrastructure, management, and cost, as both teams need to maintain different systems for essentially similar purposes—version control, reproducibility, and tracking.

3. Lack of Synergy Between Teams

The lack of integration between DevOps and MLOps pipelines also creates silos between engineering, data science, and operations teams. These silos result in poor communication, misaligned objectives, and delayed deployments. Data scientists may struggle to get their models production-ready due to the absence of consistent collaboration with software engineers and DevOps.

Moreover, because the ML models are not treated as standard software artefacts, they may bypass crucial steps of testing, security scanning, and quality assurance that are typical in a DevOps pipeline. This absence of consistency can lead to quality issues, unexpected model behavior in production, and a lack of trust between teams.

4. Deployment Challenges and Slower Iteration Cycles

The disjointed state of DevOps and MLOps also affects deployment speed and flexibility. In a traditional DevOps setting, CI/CD ensures frequent and reliable software updates. However, with ML, model deployment requires retraining, validation, and sometimes even re-architecting the integration. This mismatch results in slower iteration cycles, as each pipeline operates independently, with distinct sets of validation checks and approvals.

For instance, an engineering team might be ready to release a new feature, but if an updated ML model is needed, it might delay the release due to the separate MLOps workflow, which involves retraining and extensive testing. This leads to slower time-to-market for features that rely on machine learning components. Our State of the Union Report found organizations using our platform brought over 7 million new packages into their software supply chains in 2024, highlighting the scale and speed of development.

5. Difficulty in Maintaining Consistency and Traceability

Having separate DevOps and MLOps configurations makes it difficult to maintain a consistent approach to versioning, auditing, and traceability across the entire software system. In a typical DevOps pipeline, code changes are tracked and easily audited. In contrast, ML models have additional complexities like training data, hyperparameters, and experimentation, which often reside in separate systems with different logging mechanisms.

This lack of end-to-end traceability makes troubleshooting issues in production more complicated. For example, if a model behaves unexpectedly, tracking down whether the issue lies in the training data, model version, or a specific part of the codebase can become cumbersome without a unified pipeline.

The Case for Integration: Why Merge DevOps and MLOps?

As you can see, maintaining siloed DevOps and MLOps pipelines results in inefficiencies, redundancies, and a lack of collaboration between teams, leading to slower releases and inconsistent practices. Integrating these pipelines into a single, cohesive Software Supply Chain would help address these challenges by bringing consistency, reducing redundant work, and fostering better cross-team collaboration.

Shared End Goals of DevOps and MLOps

DevOps and MLOps share the same overarching goals: rapid delivery, automation, and reliability. Although their areas of focus differ—DevOps concentrates on traditional software development while MLOps focuses on machine learning workflows—their core objectives align in the following ways:

1.Rapid Delivery

• Both DevOps and MLOps strive to enable frequent, iterative releases to accelerate time-to-market. DevOps achieves this through the continuous integration and delivery of code changes, while MLOps aims to expedite the cycle of model development, training, and deployment.
• Rapid delivery in DevOps ensures that new software features are shipped as quickly as possible. Similarly, in MLOps, the ability to deliver updated models with improved accuracy or behaviour allows businesses to respond swiftly to changes in data or business needs.

2.Automation

• Automation is central to both practices as it reduces manual intervention and minimises the potential for human error. DevOps automates testing, building, and deploying software to ensure consistency, efficiency, and reliability.
• In MLOps, automation is equally crucial. Automating data ingestion, model training, hyperparameter tuning, and deployment allows data scientists to focus more on experimentation and improving model performance rather than dealing with repetitive tasks. Automation in MLOps also ensures reproducibility, which is critical for managing ML models in a production environment.

3.Reliability

• Both DevOps and MLOps emphasize reliability in production. DevOps uses practices like automated testing, monitoring, and infrastructure as code to maintain software stability and mitigate downtime.
• MLOps aims to maintain the reliability of deployed models, ensuring that they perform as expected in changing environments. Practices such as model monitoring, automatic retraining, and drift detection are part of MLOps that ensure the ML system stays robust and reliable over time.

Treating ML Models as Artifacts in the Software Supply Chain

In traditional DevOps, the concept of treating all software components as artefacts such as binaries, libraries, and configuration files, is well-established. These artifacts are versioned, tested, and promoted through different environments (e.g., staging, production) as part of a cohesive software supply chain. Applying the same approach to ML models can significantly streamline workflows and improve cross-functional collaboration. Here are four key benefits of treating ML models as artifacts:

1. Creates a Unified View of All Artifacts

Treating ML models as artifacts means integrating them into the same systems used for other software components, such as artifact repositories and CI/CD pipelines. This approach allows models to be versioned, tracked, and managed in the same way as code, binaries, and configurations. A unified view of all artifacts creates consistency, enhances traceability, and makes it easier to maintain control over the entire software supply chain.

For instance, versioning models alongside code means that when a new feature is released, the corresponding model version used for the feature is well-documented and reproducible. This reduces confusion, eliminates miscommunication, and allows teams to identify which versions of models and code work together seamlessly.

2. Streamlines Workflow Automation

Integrating ML models into the larger software supply chain ensures that the automation benefits seen in DevOps extend to MLOps as well. By automating the processes of training, validating, and deploying models, ML artifacts can move through a series of automated steps—from data preprocessing to final deployment—similar to the CI/CD pipelines used in traditional software delivery.

This integration means that when software engineers push a code change that affects the ML model, the same CI/CD system can trigger retraining, validation, and deployment of the model. By leveraging the existing automation infrastructure, organizations can achieve end-to-end delivery that includes all components—software and models—without adding unnecessary manual steps.

3. Enhances Collaboration Between Teams

A major challenge of maintaining separate DevOps and MLOps pipelines is the lack of cohesion between data science, engineering, and DevOps teams. Treating ML models as artifacts within the larger software supply chain fosters greater collaboration by standardizing processes and using shared tooling. When everyone uses the same infrastructure, communication improves, as there is a common understanding of how components move through development, testing, and deployment.

For example, data scientists can focus on developing high-quality models without worrying about the nuances of deployment, as the integrated pipeline will automatically take care of packaging and releasing the model artifact. Engineers, on the other hand, can treat the model as a component of the broader application, version-controlled and tested just like other parts of the software. This shared perspective enables more efficient handoffs, reduces friction between teams, and ensures alignment on project goals.

4. Improves Compliance, Security, and Governance

When models are treated as standard artifacts in the software supply chain, they can undergo the same security checks, compliance reviews, and governance protocols as other software components. DevSecOps principles—embedding security into every part of the software lifecycle—can now be extended to ML models, ensuring that they are verified, tested, and deployed in compliance with organizational security policies.

This is particularly important as models become increasingly integral to business operations. By ensuring that models are scanned for vulnerabilities, validated for quality, and governed for compliance, organizations can mitigate risks associated with deploying AI/ML in production environments.

Conclusion

Treating ML models as artifacts within the larger software supply chain transforms the traditional approach of separating DevOps and MLOps into a unified, cohesive process. This integration streamlines workflows by leveraging existing CI/CD pipelines for all artifacts, enhances collaboration by standardizing processes and infrastructure, and ensures that both code and models meet the same standards for quality, reliability, and security. As organizations race to deploy more software and models, we need holistic governance.

Currently, only 60% of companies have full visibility into software provenance in production. By combining DevOps and MLOps into a single Software Supply Chain, organizations can better achieve their shared goals of rapid delivery, automation, and reliability, creating an efficient and secure environment for building, testing, and deploying the entire spectrum of software, from application code to machine learning models.

We've compiled a list of the best IT infrastructure management services.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

A dual U.S.-German citizen has been arrested on charges that he traveled to Israel and attempted to firebomb the branch office of the U.S. Embassy in Tel Aviv, officials said Sunday.

(Image credit: Ohad Zwigenberg)

Grant Hardin was the police chief of Gateway, Ark. for about four months in 2016. Corrections officials did not provide any details about how he escaped.

(Image credit: AP)

President Donald Trump said Sunday that the U.S. will delay implementation of a 50% tariff on goods from the European Union from June 1 until July 9 to buy time for negotiations with the bloc.

(Image credit: Manuel Balce Ceneta)

Hints and answers for the NYT Connections: Sports Edition puzzle, No. 245, for May 26.

Hints and answers for Connections for May 26, #715.

Here are the answers for The New York Times Mini Crossword for May 26.