The 2nd time I felt like a leader was during my 1st week with a new client. Since my time as an in-house Airflow expert, I’ve found a place where I can be responsible for both architecting and building things. I was a part of a fantastic team. We were responsible for doing data integrations for their internal workflow engine. Imagine zapier, a workflow engine, with drag and drop and everything but tuned for a different usecase
Before I joined, the company had made an architectural choice that was ready for me to pick up: To use Airbyte for data/app integrations. I was tasked to get familiar with the work that had gone into it, the current workflow engine and previous Airbyte documents.
I started with a clean slate. I worked through the existing workflow specifications, existing data integrations, and why it was not working, and finally came to Airbyte. I took one look at it, and I immediately knew it wasn’t the right move. A decision that the CTO approved, but I wasn’t in the game.
I had my reasons
Airbyte was young, and its open-source API support and documentation wasn’t there yet, integrating it with our workflow engines would’ve been nightmare.
I did integrate a few POCs. The integrations specification, for example, if you’re connecting Google Docs, our system needs to construct JSON at least two pages long that Airbyte can understand. All our engineers need to be masters of this spec to debug if any workflow fails, which means we are creating high knowledge and code coupling
Finally, given the team's size and expertise, creating an in-house SDK to tackle future integrations would be a good use of our resources.
I mapped this out in a new Feasibility analysis document, which includes POC code to support my claim.
I walked up to my immediate manager about my claims, and he was a little shocked to see that something that went through multiple design approvals could not hold up. But it was already CTO-approved. I don’t remember exactly what happened, but I asked my manager to support me in the next Design review meeting. I told him I was hired to do exactly this.
Now comes the hard part. I had to present it. The Zoom room started filling up, 19, 27, 35, 43…. Almost all engineers from all teams were there. There I was, Bhavani Ravi, someone who joined just one week back, taking on against a technological choice that a whole company was sold on.
I walked them through how we arrived there, what I’ve learned in the last week, emphasised that I’m still learning, and opened the doors for feedback. Then, I unleashed the POCs and showed the whole team the 1.5 pages of JSON that did it. There were questions, and I had answers. What’s the alternative? In-house SDK + Integrations. Timeline?
Four engineers shipped 22 integrations to production in the next two months while constantly upgrading the SDK. The SDK had better error handling, monitoring, and retry. It became an integration powerhouse. Every new integration had a two-week development lifecycle: one week to understand the external APIs and one week for the integration.
The second time I felt like a leader was during my first week with a new client when I asserted my expertise against a team of 60 engineers, as a newbie on the team, to stand up for something I strongly believed in. I would go on to create more such splashes, be it proposing a full-on data transformation engine or finding my way into devops tasks, even though it wasn’t a part of the “JD”
But that wasn’t the part that created the lasting impact. It was the part where my knowledge was received with intellectual curiosity rather than ego(a common theme in many engineering teams even today). That set a standard for the team I would lead later: to create a culture of safety where we debate ideas, not people or their hierarchy. The groundwork for my leadership was reinforced there: Ownership, Autonomy, Meritocracy, and continuous Learning—a culture not just on paper but a practice in how people show up.
I didn’t know I was preparing for something bigger, all the while taking notes of my managers’ leadership style. My time to take the wheel came sooner than I expected, and it was time to put the learnings into practice