An engineer’s journey from the frontlines of o11y to the forefront with Agentic AI

Here at Resolve AI, we're always excited to welcome new talent. Especially when they bring a wealth of experience from renowned companies, we love to understand how the transition feels.

So, when Mehul, with his deep background in Observability from Splunk and DataOps from Astronomer, decided to join us, I (Juan) was keen to understand his perspective. We all join startups with an expectation: immense energy, rapid innovation, and a touch of the unexpected. I wanted to hear firsthand from Mehul: what were his genuine first impressions? Did the reality of joining Resolve AI align with his expectations, or did he run into any surprises?

[Juan] Mehul, you've spent the last few years at Splunk and Astronomer. What about Resolve AI's mission resonated with you?

[Mehul] It’s simple, really. I've lived through the pain. I’ve spent the first few years of my engineering career in Observability, and then in DataOps, I saw firsthand the immense pressure and complexities that on-call engineers, software engineers, data engineers grapple with.

In fact, this problem was so acute and visible within our customer base at Astronomer that last year, I teamed up with some colleagues for Cohere’s Enterprise AI Hackathon. We built a project focused on “Self-Healing Pipelines”, and it actually won first place. That experience crystallized for me just how the need for better operational solutions is so critical and widespread.

And what makes it so incredibly tough is that you're dealing with these vast, interconnected systems where the complexity is so high for a single engineer or even a team to completely understand. In these scenarios, tribal knowledge often becomes a bottleneck and is locked in the heads of a few seasoned engineers. I found that traditional tools can be too brittle to solve this problem. While newer AI-centric approaches like chatbots or workflows show promise, I found they are limited to assisting with isolated tasks. Solving system complexity requires deep understanding and built-in reasoning to autonomously investigate and pursue novel issues.

So, when I learned about Resolve AI's mission to bring Agentic AI to Software Engineering, it felt like an incredibly natural fit. It wasn’t just another interesting problem; it was THE problem I was already passionate about solving.

[Juan] Joining an early-stage startup can be a big shift. What were your expectations, especially coming from larger, more established companies?

[Mehul] Honestly, this is the first time I am working at an early stage company. So, I was prepared. You hear stories, right? I was looking forward to the rapid pace but was also anticipating uncertainty in tooling and processes. Also I was bracing for a generally higher level of uncertainty from my day-to-day. I figured that's often the trade-off for being at the ground floor of something potentially ground-breaking.

[Juan] And what was the reality when you walked through the doors at Resolve?

[Mehul] 2 weeks in, I can confidently say my initial expectations were completely off-base – in the best way possible. One of the first things that struck me was the level of investment in developer tooling. It's impressive. We can spin up the entire system locally on our laptops with a single command. For a new engineer, that’s gold; it meant I could start learning the system and genuinely start contributing within hours, not days or weeks. The CI/CD pipelines are streamlined, fast, and reliable, which breeds confidence when you’re shipping code. And keeping up with the latest methodologies, AI-assisted coding practices are seamlessly integrated into our IDEs and daily workflows. This tangibly helps us move faster and allows us to dedicate more of our brainpower to solving truly hard problems.

[Juan] Given that Resolve AI was founded by co-creators of OpenTelemetry, a strong observability culture might be expected. What did you find?

[Mehul] The focus was expected, yes. But I was still incredibly impressed by the depth of its integration. Observability isn't an afterthought here; it's foundational to how engineers build, debug, and operate. Whether it's during our weekly operations reviews, product demos, or even when writing the 'definition of done' for a task, the focus on logs, metrics, and traces is constant. This commitment creates a healthy culture of accountability and meticulous attention to detail. Often these things are crucial for system stability and understanding.

[Juan] Fast-moving startups often prioritize speed to market. What practices did you notice at Resolve AI?

[Mehul] That's another area that has been a really positive discovery. Our team is striking a very thoughtful balance between the need for speed and the wisdom of long-term scalability. Shipping quickly here doesn't mean cutting corners on quality. Technical debt is actively discussed and tackled head-on. Solutions are architected with future growth in mind, not just immediate needs. And importantly, production incidents, when they occur, are treated as invaluable learning opportunities. There's no blame game, just a collective effort to understand and improve. We also 'dogfood' our own product extensively, which is critical. It makes sure we're not just building tools we think engineering teams need, but tools that we as an engineering team find indispensable.

[Juan] It sounds like you’ve hit the ground running. Have you been able to make an impact already in these initial weeks?

[Mehul] Absolutely, and that’s been incredibly rewarding. The environment is set up for engineers like me to hit the ground running quickly. In just my first two weeks, I've been able to implement a change that reduced our database size growth by half, which has direct cost-saving implications. I’ve also worked on resolving some operational issues in the RPC layer between customer environments and our control plane. Plus, I got to ship a new product feature to improve how users interact with Resolve AI via Slack.

A big part of this rapid onboarding and impact is the collaborative, in-person office culture. This was actually one of the things that really drew me to Resolve AI. The ability to have ad-hoc product brainstorming sessions, to pair program on a tricky piece of code, or to hash out design decisions quickly and informally, is incredibly powerful. It allows for quick iteration and effective decision-making. And, of course, the daily lunches and Friday happy hours are a fantastic way to connect with colleagues and turn coworkers into friends.

[Juan] Looking ahead, what excites you most about your journey with Resolve AI?

[Mehul] It’s been an incredibly fun, enlightening, and frankly, validating start. I'm genuinely excited to continue building alongside this team. We're working on solving deeply challenging problems that can fundamentally improve engineering operations for teams all over the world. The combination of a compelling mission, a surprisingly mature and supportive engineering environment, and a fantastic team culture makes me very optimistic and eager for what's next."

About Resolve AI

Resolve AI is the agentic AI company for software engineering founded by the co-creators of OpenTelemetry. By combining our deep expertise in building developer tools and observability with state-of-the-art agentic AI, our mission is to increase engineering velocity by transforming the way engineers build, deploy, and maintain real-world software systems.

Resolve AI autonomously troubleshoots and resolves production issues, freeing up engineers to focus on building. Our agentic AI understands your production environments, reasons like your seasoned engineers, and learns from every interaction to give your engineering teams decisive control over on-call incidents with autonomous investigations and clear resolution guidance.

With Resolve AI, customers like Datastax, Tubi, and Rappi, have increased engineering velocity and systems reliability by putting machines on-call for humans and letting engineers just code. Interested in learning more about our Agentic AI approach to production systems? Say hello.