Alteryx CTO DV Lamba on Why AI Will Create Chaos in Software Development - And Why Platform Engineers Should Embrace It

By Arshad Sayyad, Co-Founder and CBO of StackGen

The landscape of software engineering is evolving rapidly, with platform engineering emerging as a critical discipline for modern organizations. Recently, I had the pleasure of sitting down with Digvijay "DV" Lamba, Chief Technology Officer at Alteryx, for an insightful discussion on the transformation of platform engineering, the challenges facing engineering leaders, and the profound impact of AI on software development practices.

DV brings a wealth of experience from his impressive career journey. Currently overseeing a global team of 650 engineers at Alteryx, he joined the company through the acquisition of his startup, Lore IO. His background spans successful entrepreneurship and leadership roles at major companies including Walmart, giving him a unique perspective on engineering organizations across different scales and stages.

In our conversation, we explored how platform engineering is evolving into a product discipline, the shifting balance between centralization and autonomy, and how AI is fundamentally reshaping the engineering landscape. Below are key insights from our discussion that will resonate with platform engineers, engineering leaders, and CTOs navigating these transformative times.

Q&A with DV Lamba, CTO of Alteryx

Arshad: You've called it "modern engineering" because things are changing rapidly. What are some of the big macro disruptive trends, particularly around platform engineering?

DV: I think the main thing that has changed is that engineering leaders need to be operational leaders, not just delivery leaders. You can't just deliver features - you need to care about what you're building end-to-end, have business outcomes in mind, and understand the 360-degree impact of anything you do.

When it comes to platform engineering specifically, it's really important to think of platform as a product, with internal teams as your customers. When you start thinking of these teams as customers, having a roadmap, and measuring success in terms of the speed, velocity, and experience you're providing to the people using your platform, your approach changes fundamentally.

One key challenge is finding the right balance: how do you make teams that depend on your platform more self-service so they're not bottlenecked by you, while also providing the right amount of governance so it's not complete chaos? This balance between centralizing everything into the platform team versus giving more control to application teams is different for every company.

Additionally, with DORA metrics and other tools, you can be much more measured and outcome-based in how you run and build your infrastructure and platform. Everything needs to be more automated - it's not just continuous integration and CI/CD for code, but CI/CD for everything. Infrastructure as code, continuous deployment, faster feedback loops, higher velocity, continuous releases across the board - how do you enable the same ethos that has been built across the application stack deeper into the platform?

Arshad: You talked about being operational leaders. Everyone's excited about dramatic velocity increases in software development lifecycles thanks to AI copilots, but what are the implications for the rest of the value chain - CICD, DevOps, DevSecOps, platform engineering, deployment cycles?

DV: There are a couple of important dynamics happening. The teams on the left that are building applications are increasingly owning bigger portions of the stack. When you talk about embedded DevOps, embedded DevSecOps, quality within engineering - the engineering team becomes responsible for everything: reliability, deployment, everything in their own stack.

AI has also made the skill gap smaller, enabling teams to go deeper and do more across multiple clouds. This empowers those teams, gives them autonomy, and makes them move faster. But when you have 10 of those teams, or in Alteryx's case 60 teams, the challenge for platform and infrastructure engineering becomes: how do you build governance without slowing them down?

The key is finding ways to use AI or other mechanisms to provide golden paths or pre-baked, governed components that teams can use as-is, while at the same time providing the right level of configuration options. In some cases, you also need to provide escape hatches where teams can extend what you're building for specific use cases.

The more sophisticated your platform is, the more you'll provide all of these at once: golden paths, governance for teams using escape hatches, and the right depth of configuration options. If you provide too many options, you increase complexity for those teams and thus require more skills. That's the balance we discussed.

Companies like StackGen are trying to provide that core infrastructure that helps platform teams build the right balance. This is a hard thing to build - it's complex with service meshes, security layers, and many other components. The required skill set is difficult to find, so having the right tooling and platform is essential.

Arshad: StackGen's vision is that developers should eventually be able to deploy on their own, when they want, on any cloud. But achieving this requires change management, behavioral adaptation, leveraging AI, and the right ecosystem of integrated tools. Having built companies and led transformations, how do you see this playing out over the next year or two? What are the biggest challenges for platform engineering teams?

DV: Let's talk about AI's role first. There's both automation with AI, allowing you to do more with fewer people, and a fundamental shift in how platform teams operate.

When you start thinking of platform as a product, you're no longer just a ticket organization servicing requests and becoming a bottleneck. You're an enablement organization pushing work to self-service engineering teams. It's almost like selling a product - you go to a company, try to sell to them, and then work to get adoption.

To get adoption, you need to measure NPS scores and user experience. You also need to do the right kind of change management - you don't get everything inbound from the teams using your platform. You need to proactively influence them and drive change. This non-core part of collaboration and communication is where AI will play a big role that not enough people consider.

For example, when you make a change in your platform, enable a new policy, or become PCI compliant, you have to change processes and how teams interact with you. With many development teams, how do you drive that change, upgrade, and adoption? As they have AI agents on their side, I believe the communication and rollout of new capabilities will become much more automated and faster.

The challenge for platform teams is that you suddenly go from being focused on building your own things to becoming an influence organization where perhaps 60% of your job is building, but 40% is influencing the rest of the organization on modern practices. It's the whole concept of guilds or centers of excellence. Platform teams will be a combination of a software team and a center of excellence, requiring different skill sets from leaders and different capabilities in tools.

Arshad: We often see matrix structures in companies with centralized teams and decentralized business teams facing the market. How do you see this evolving, and what would you advise teams in these structures?

DV: First, there's no single right answer, even for a single company - the answer changes over time. It's different for a 100-person company versus a 2,000-person company, different with one product versus five products, and changes after acquisitions.

As a platform team, you have to build a platform that can provide different levels of flexibility between centralization and decentralization for different customers. Some teams might be less sophisticated and need more centralization, while others might be more sophisticated. Some may be working on mature products, others on early-stage products.

The important thing is to build a platform where any consuming team can self-serve, self-understand, and self-regulate the right balance, while the central team maintains governance through metrics and usage data. This lets you identify which teams need the most support because you're doing most of the work for them, versus teams that need a different kind of support around developer experience because they're making many changes themselves.

From a tooling perspective, you need both the right configuration options for different teams and the right people and processes to handle the complexity. The more sophisticated and larger your organization, the more complexity exists. Anything that can simplify that complexity will help tremendously.

The world is moving toward continuous delivery and releases across the platform - infrastructure and security fixes going out continuously, all the time, separately for each team with independent deployments. In this continuous world, different teams are going deeper into the stack and making changes in different ways. That's what a sophisticated platform looks like.

Most companies won't reach that state immediately - they have to pick a golden line and then try to vary by maybe 10%. If you can vary by 30%, you're already ahead. The flexibility of your platform is how I would measure its sophistication.

Arshad: For platform companies or SaaS companies, one of the biggest challenges is deploying across multiple clouds, especially with data residency requirements. At Alteryx, you've solved many of these problems. Could you share your approach?

DV: Even at Alteryx, we've been on a journey where we've tried different approaches and learned from them. We started with a more centralized model where a few teams focused on getting our software to deploy across multiple clouds. We had different repositories for each cloud with infrastructure code for each.

As more applications started deploying to multiple clouds, that complexity got exposed to those teams. There's a continuous debate around how much abstraction we should provide to simplify things for teams (taking control centrally) versus letting teams self-serve and go deeper.

What we find is that more sophisticated teams want the configuration options. They have the DevOps knowledge and get frustrated when blocked by a golden version of the platform that restricts them. Other teams appreciate the guardrails because they aren't focused on that kind of complexity.

At Alteryx, we've reached a point where many teams have started getting deeper into DevOps and platform work. Our infrastructure team is actively working on providing deeper self-service and different balances for different teams. We're also addressing the multi-cloud problem, where as you go deeper into the stack, you have to do everything multiple times for different cloud vendors.

Alteryx operates with a core control plane on one cloud, and data assets running in whichever cloud vendor the customer is using. We manage hundreds of instances across customers, which makes for a complex setup. We're still finding the right balance.

Arshad: As you think about the journey for platform engineering over the next 2-3 years, what excites you about AI and what worries you?

DV: I'm a big believer in AI. I think about how it's transforming how we build software and do engineering across all teams, including platform teams.

There are two levels of what you can do with AI. First, you can look at what you're doing today and ask if you can do it easier, faster, and better with AI. There are many opportunities in both creating code and improving the entire SDLC process, especially around communication, approval processes, and automation.

But where I'm spending most of my time is thinking about what happens after those initial improvements. Once you've applied AI, the nature of engineers and their impact fundamentally changes. AI increases the complexity of what a junior engineer can do and the diversity of skills they have.

I call this the ability to create "10X engineers." Every company has these engineers who can dive into anything, understand the architecture, and excel. Now imagine taking every junior engineer and making them a 10X engineer who understands the architecture, all the code, and best practices. What happens when you have 200 10X engineers instead of 10 10X engineers and 190 great engineers?

This requires changing how we think about software development. The very existence of platform teams stems from the belief that doing things in one place is better than having everyone build their own solution. With AI, perhaps a little more chaos is acceptable, though it will also result in a bigger governance load.

I think we're moving toward more chaos, more duplication across teams, and more variety in what's being built. To truly leverage AI, you have to embrace this and say teams will do things in many different ways.

The challenge for platform and central teams responsible for quality and governance is using AI to make sense of the chaos, find outliers, and identify exceptions. Instead of hard-coding governance rules, you can use AI to understand what teams are doing and help them get to the right outcomes when needed. This is behavioral and organizational change in ways we haven't thought about before, and it's the most exciting part of the next few years.

Arshad: We've been measuring productivity improvements from AI copilots with our engineering team at StackGen. After standardizing on Cursor about five months ago, we're seeing about 41% more commits in just three and a half months. Developers report 40-50% productivity increases. This clearly connects with what you're saying. Changing gears - what are your passions outside of work?

DV: I have young kids right now, so that takes up a lot of my time. We love to travel as a family - we were just in Peru, visiting the mountains and the Amazon jungle. With kids taking up most of my time, I've found a few small hobbies of my own. I've always been into reading and video games, though I don't get much time to finish things. I travel a lot for work, so I try to catch up on reading or play some games during those trips.

Arshad: If you were to recommend two or three good books from your experiences and readings, which ones would you suggest?

DV: I'm currently reading "Seven Powers," which explores structural sources of competitive advantage for companies. It categorizes different ways that companies build moats and includes fascinating stories about companies like Netflix.

I'm also reading "The Dawn of Everything," one of those broad-scale history books that cover everything from prehistory to today. It's an academically rigorous take on the history of thought, freedom, and human evolution, which is particularly interesting given everything happening in the world today.

Finally, on the fun side, I read a lot of science fiction and fantasy. I recently started "Dungeon Crawler," a multi-book series that's taken social media by storm. It's a unique book with a cat as one of the main characters - unlike anything I've read before, but there's a reason it's so popular.

Arshad Sayyad is the Co-Founder and Chief Business Officer of StackGen, a platform revolutionizing how technical teams build and deploy modern applications. With over 25 years of technology leadership experience, Arshad has led digital transformation at some of the world's largest enterprises.

Digvijay "DV" Lamba is the Chief Technology Officer at Alteryx, leading a global team of 650 engineers. He joined Alteryx through the acquisition of his startup Lore IO and previously drove platform initiatives at Walmart.