harisrid Tech News

Behavioral Interviewer : “Tell me about a time you had to deal with limited information and you made a wrong decision.”

A Primer

It’s a stumpy question for senior-level and above behavioral interviews; I’m still determining the most appropriate manner to story tell. But here I go 🙂 !!!

Situation

I’m working as a senior engineer at Geico where I’m taking the lead on an organizational challenge problem : achieve 100% compliance coverage and severely reduce regulatory risk for PII data across GEICO’s enterprise data platforms.

The stakes were high – – I have to meet an aggressive six month regulatory timeline; if unmet, the company would have to cease New York operations, resulting in millions in estimated annual losses.

But it’s also incredibly ambiguous! I’m limited.

I’m limited in what I know, what data has the PII, and what the best long-term architecture should even resemble, given enumerable systems.

Task

Albeit the uncertainties, I had to make a strategic call. So I set about a task : I developed out a rules engine capable of sensitive data classification. And I made two key assumptions.

1. Firstly, that the rules are constant.
2. And secondly, that the data sources are constant.

By doing so, I could quickly achieve the compliance milestone with a working prototype.

Actions

But even though I delivered the first part of the ask in six months, I ran into unexpected issues later when extending the application cross-functionally. My assumptions broke. Workflows that worked with one input set immediately failed on alternative input sets, because the underlying rules and the datasets changed.

Consequentially, I had to take a few actions. I had to delay future feature extensions, partner with engineering leads and compliance leads, and rearchitect the application for modularity and extensibility. This effort took about 2-3 weeks, but led to the scaling of tool adoption from team level to company level.

Results

Now despite these refactor efforts, I was still able to deliver the 100% compliance coverage ask while keeping systems scaleable and adaptable, even after the initial 6 month ask. I also succeed in setting the foundation of internal data governance infrastructure.

Key Takeaways and Learnings

If there’s anything I learnt from this experience, it’s to strongly invest upfront time in documenting key assumptions. From now on, I always flag these assumptions as “critical risk areas”; I design with a few caveats and extensions in mind to minimize future organizational pain points.