If you’re the kind of engineer who likes the frontier—where the spec is still evolving and the “right way” is something you figure out together—this one’s for you. In this chat, Ben and Sirish talk about what it’s been like building with emerging AI tech: pairing complementary strengths, keeping feedback loops tight with quick “let’s sync?” moments, and using modern AI tools to ramp up fast and ship with confidence (even while the ground keeps moving). 

Tell us a bit about yourself 

Ben: I joined Veeam in October 2024. Over the past 8 years I've built my career as a backend software engineer, working across a range of companies and startups. That variety has given me broad exposure to different engineering cultures and challenges, with a particular focus on distributed systems. When I heard about the AI backend engineering position at Veeam, I jumped at the opportunity. 

Sirish: My career has really followed the evolution of storage. I started out at legacy storage companies, moved into a cloud storage startup, and joined Veeam in 2020. Looking back, each step tracked the industry forward - from traditional on-prem systems, to cloud-native, and now into AI-powered data protection. After a few years deep in cloud and Kubernetes, I was ready for something new. I started noticing what AI was becoming capable of and got genuinely curious - both about the technology and about how I could grow with it. The timing aligned perfectly with a need on the Veeam Data Cloud Intelligence team, and when I looked at the work, it clicked immediately. I'd spent my career building interfaces - CLIs, APIs - the layers that connect systems to the people and tools that use them. MCP, the Model Context Protocol, is essentially that same idea but for AI. It felt like a natural next step, and exactly the kind of thing I'd want to build. 

What inspired you to take on the challenge of working with cutting-edge AI technology? 

Ben: When I joined, I immediately got to work on the MCP project, which was still relatively young, meaning there was real opportunity to help shape its direction rather than just execute on a plan already set in stone. That was exciting. Sirish, my partner on the MCP team, was incredibly welcoming and got me up to speed quickly, which made a big difference in those early weeks. Honestly though, when I started I was still a bit skeptical about AI, particularly around its capabilities in coding and what a truly good AI product would even look like. In some ways, the MCP project turned out to be the perfect environment to work through that skepticism. We built something I stand behind, and along the way I came to see just how powerful these tools can be when they're used well within a strong team. 

Sirish: What made it especially exciting was that AI itself became a tool in my onboarding. Getting up to speed on an unfamiliar codebase can be slow and frustrating, but being able to use AI to explore, ask questions, and understand the architecture quickly was a game changer. It felt like the technology I was building was also helping me build it. Being a greenfield team, there were naturally a lot of unknowns at the start. No established playbook, no well-worn patterns to follow. That uncertainty was real - especially when the technology itself was also moving fast. But the flip side was freedom. We got to decide how we wanted to build, how we wanted to work together, what standards we wanted to set. The uncertainty and the freedom came as a package deal, and I came to really value that. 

When did this stop being an idea and become a real, urgent problem to solve? 

Ben: There was a specific moment that really crystallized it for me. With an MCP server, you're essentially building an API for AI agents, and the work is very granular at first. You're building one tool at a time, carefully perfecting each individual tool call. We had implemented a suite of tools that allow agents to navigate a subset of our public API, and one day while testing I decided to throw it a random, open-ended question: "Analyze my workloads for any suspicious activity." And it just did it. It "reasoned" about the API, chained together the necessary tool calls on its own, and synthesized a report flagging things like suspicious login activity, unusual policy changes, etc. The technology is evolving so fast that you're constantly discovering what it's actually capable of. 

Sirish: That was the moment it stopped being an architectural concept and became something real. I remember thinking - this is why we're building this. 

How did you work together? 

Ben: Sirish and I turned out to have naturally complementary skill sets, which made the ownership split almost immediately obvious. Sirish started the MCP server before I joined, and when I came on board I took over the Python implementation and drove it to production, building out most of the tools, the security layer, testing, and the middleware layer. The infrastructure side of making this a robust, reliable service was much more Sirish's wheelhouse, so we each gravitated toward what we do best. But we were always willing to jump on a call and bounce ideas off each other when it was useful. That balance of autonomy and openness is what made it work so well. Day to day we worked mostly independently, trusting each other to get our respective pieces done. 

Sirish: We complemented each other really naturally - I gravitated toward architecture and infrastructure, while Benjamin, coming from a strong Python background, took ownership of building out the MCP service and tooling. Early on, I built an initial version of the MCP service and it was pretty messy - Python isn't my strongest language, and it showed. Benjamin came in and cleaned it up without making a big deal of it. That kind of quiet, generous collaboration is hard to find. On the flip side, he regularly reaches out to me when platform changes come up. It goes both ways. There's no ego about who owns what - just a shared interest in getting it right. In practice, neither of us felt like our responsibility stopped at our own boundaries. If something needed attention, we both felt accountable for it - regardless of who built it. And great collaboration often looked like two words: "let's sync?"  

What does it feel like to work on such new technology? 

Ben: Coming in, I had never even heard of the MCP protocol, so getting up to speed on that was one of my first challenges. Sirish was great about always being available to jump on a call and walk me through things, which helped a lot in those early weeks. But honestly, the biggest accelerant for onboarding was the AI coding tools that the Intelligence team was already championing internally. The Veeam codebase is spread across several repos, which can be daunting at first, but with tools like Warp it was no problem. You can practically have a conversation with the codebase, asking things like "where is this functionality implemented?", "why was this design chosen?", "how do these two services interact?" It made getting productive fast so much more achievable. 

Sirish: Working on technology this new feels great, but you can't help feeling a little unsettled knowing things are going to change rapidly. The spec you built against today might look different in six months. So there's a constant awareness that you have to stay nimble and not get too attached to the way things are right now.  

How did this project change you as an engineer? 

Ben: The biggest shift was in how I think about AI. I came in skeptical, and working closely with this technology every day helped me see how powerful of a tool it is. Beyond that, working with fast-moving technology taught me to get comfortable with uncertainty. The MCP protocol was young when I joined, best practices were still forming, and things were changing quickly. Learning to make good decisions without a stable floor underneath you is something I'll carry into everything I do next. 

Sirish: Working in this space has amplified something that's always been part of how I operate - I move fast and like to get things done. But AI has become a genuine accelerant for that. The biggest mindset shift has been learning to trust it more. Early on you're skeptical, you double-check everything. But the more you work with it, the more you develop an intuition for when to lean on it and when to verify. Engineers who learn to work fluidly with AI are just going to operate at a different level - and I feel like I'm starting to get there. 

Why should an engineer join Veeam? 

Ben: In my experience with Veeam so far, the team is wonderful, supportive, friendly, and everyone here has a real drive to ship. There's a genuine sense of everyone pulling in the same direction, working together to bring ideas to fruition. 

Sirish: As AI grows, data protection and security become more critical, not less. We're building the tools that make that possible, at the intersection of AI and data protection, which is one of the most interesting places to be right now. You get great tools, genuinely hard problems, a strong team, and a remote-friendly environment. For an engineer who wants to build things that matter and move fast, I think Veeam is worth a serious look. 

What makes the work meaningful to you? 

Ben: Honestly, it comes down to the people. Building alongside coworkers who care about what they're making creates a dynamism that's hard to put into words but easy to feel. That energy is what keeps me motivated and makes the work feel worthwhile. 

Sirish: What makes the work meaningful to me comes down to two things: being on the frontier, where the answers aren't all written down yet and you're learning constantly - and ownership. When you can point to something and say "I built that, I made those decisions" - that's deeply satisfying. Not in an ego way, but in the sense that your work has a real identity and a real impact. 

What do you like to do in your free time? 

Ben: I'm a big movie watcher. I also co-run a local book club — we recently finished Moby-Dick, which was quite the undertaking. And like most Bay Area people I try to get out into California nature as much as I can. My partner and I are actually heading down to Big Sur next month, which I'm really looking forward to. 

Sirish: Outside of work I stay pretty busy. Golf is a big one - there's something about the combination of focus and being outdoors that I really enjoy. I do CrossFit regularly, which is a good counterbalance to sitting at a desk. I love cooking, watching movies, and reading. And lately I've been trying to get into photography - still very much a beginner, but it's been a fun new creative outlet to explore. 

Want to build on the AI frontier? 

Looking for software engineering work that sits at the intersection of AI, developer tooling, security, and data protection? Explore our engineering roles and tech careers, and join the team building real-world AI systems that ship.