What is the most innovative idea you have ever had? How did you come up with it and how did you go about implementing it?

Question

Anonymous · Accepted Answer

SITUATION: AT DAYFORWARD, OUR MANUAL UNDERWRITING PROCESS FOR POLICY APPROVALS
WAS SLOW, ERROR-PRONE, AND INEFFICIENT. THIS RESULTED IN DELAYS AND
INACCURACIES, WHICH AFFECTED OUR OVERALL OPERATIONAL EFFICIENCY AND CUSTOMER
SATISFACTION.

Task: My task was to develop a solution to automate the policy approval process,
replacing the manual tasks with an automated system to streamline operations and
improve accuracy.
Action:
 * Research & Design: I began by analyzing the existing manual process to
   identify key areas for automation. I worked closely with the underwriting
   team to understand their criteria for policy approvals and designed an
   automated system using machine learning algorithms.
 * Development: I developed the system using Python, integrating it with our
   policy management platform. The system included modules for data extraction,
   risk assessment, and decision-making based on historical data.
 * Testing & Deployment: I rigorously tested the system with historical data and
   conducted user acceptance testing with the underwriting team. After
   successful validation, I implemented the system in phases, starting with a
   pilot program to gather feedback and refine the solution.
 * Monitoring & Improvement: I continuously monitored the system’s performance
   and made iterative improvements based on feedback and real-time data.

Result: The automated policy approval system reduced processing times by 50% and
significantly decreased errors associated with manual underwriting. This
enhancement improved operational efficiency and allowed the underwriting team to
focus on more complex cases, leading to higher productivity and increased
customer satisfaction.
This project was a prime example of leveraging innovative technology to solve
real business problems and achieve substantial improvements in both operational
efficiency and service quality.

Anonymous · Answer

Situation: When I started working on a project to improve the bulk job
framework, I examined the existing issues and reviewed the codebase to
understand the current setup. I noticed that our system used a basic Redis
set-based concurrency limiter, which was causing challenges in handling expired
or stuck jobs, as well as in tracking job durations accurately.
Task: My task was to identify a more effective concurrency limiting approach
that could scale with our job volume and provide better control over job
timeouts, as well as efficient cleanup capabilities.
Action: During my research, I recalled reading about Redis sorted set-based
concurrency limiters, which seemed like a promising solution for our needs. I
explored the sorted set approach further, analyzing how it would fit our scale
and traffic patterns. I outlined the advantages of a sorted set-based limiter
over the existing set-based approach, including improved tracking of job
durations, granular control over job timeouts, and more efficient querying and
cleanup of old jobs. I presented this idea to the team, explaining how it would
address our current limitations.
Result: The team accepted my proposal, and we implemented the sorted set-based
concurrency limiter. This approach proved highly effective, enhancing
scalability and allowing for more accurate job tracking and better control over
job timeouts. It also streamlined the cleanup of expired jobs, significantly
improving our system’s reliability and performance. Furthermore, this
concurrency limiter was adapted for use on the app server to control excessive
API traffic, and it is now available across the application to manage traffic in
other areas as needed, making it a versatile tool for resource management.

Anonymous · Answer

In a project where I was part of building an event-driven messaging system for
our company, we encountered a significant challenge reported by our customers.
They needed a way to track the status of messages they published, including
whether a message was archived in S3, routed to specific queues, or successfully
consumed by downstream applications. This end-to-end tracking was crucial for
ensuring transparency and reliability in message delivery.

My task was to design a solution that would allow customers to monitor the
status of their messages from ingestion to delivery in near real-time.
Additionally, we needed to address the issue that when customers requested
status updates or needed debugging, we had to manually check logs, which was
time-consuming and inefficient.

I proposed a design where each part of the system that handled the message would
update its status at important stages to AWS Kinesis stream. We then implemented
a consumer service that processes these status updates from the stream and
stores them in a database optimized for querying. This setup allowed our
customers to query the status of any message, and our microservice would return
the relevant status updates, providing a clear view of the message’s journey
through the system.

This solution was accepted and successfully implemented. It proved to be highly
scalable, capable of handling large volumes of status updates without impacting
system performance. Customers now have the ability to check message statuses in
near real-time, greatly enhancing their ability to monitor and trust the
system’s end-to-end delivery process. Moreover, the solution significantly
reduced the time and effort required for debugging, as customers could now
access status information directly, without needing our team to manually check
logs. This not only improved customer satisfaction but also allowed our team to
focus on more strategic tasks.

Anonymous · Answer

Initially when I joined the project, my manager asked about whether the given
testcase was automated. Since, there was no documentation, I had to go over the
regression suite which consisted of almost 2000 testcases and find out if it was
automated. This was a tedious task and the challenge here deep dive into the
code and find out the testcase which took couple of hours.
So to overcome this challenge, I came up an idea to document the testcases at
one place so whenever someone has to learn or get information about the
regression can look at the document and easily find out the details.
For this, I started by consolidating testcases from code and their existence on
test case management tool, and bifurcate these testcases based on the high level
scenario and application. Once testcases were bifurcated, I created separate
issue links and documented. I also created a separate table which included no of
testcases for a particular scenario. This helped to standardize the QA process
by maintaining high quality standards and save lot of time when someone asks
about particular testcase, or when automation had to be updated, or for new
comer it is a place to start consolidating details from.
Thus after setting up the processes, it helped in standardizing QA practices and
I was also appreciated by the time as making the process simple helped to save
time and efforts going forward.
After this, we documented other critical processes and flows as well which
helped the team to learn and save critical time, and maintain high quality
standards for QA process.
Thus from this incident I learnt how setting up the process and maintaining high
quality standards helps in saving crucial time.

Anonymous · Answer

When I joined my new team , I took lead on a new project which was complex in
nature, had many unknowns and our team had to work on new framework, techstack,
data models. The team was following agile methodology and will start with the
work in first sprint. there was no concept of zero sprint or story grooming
prior to the start of first sprint.
with the existing approach I envisioned some issues:
1. Because the number of unknown were high, the chances of doing wrong
estimations for development were high.
2. There were high chances that the prioritisation is not right and we might
have to revisit the features again and again.
3.  We might find out about unknowns while we are in the development process and
that might cause delays.
To overcome these issues, I suggested some changes in our approach:
1. Increase the planning for one more release cycle and group the features as
now, next and later.
2. For the features planned in the now group , prioritise them for the next
release and complete the research and PRDs for them one sprint early. That means
dev and QA will have 2 weeks extra to go through the stories, find out the
dependencies and share the feedback. This helped in identifying if we need POCs
for any feature or we can right away start developing them. Or if further
modularisation is required.
3. I suggested to track the features related to new project in a different
group, because to begin with the pace of the delivery will be slow.
When we begin, we took one complete release only in the RnD and understanding
the new framework and doing small POCs to validate our understanding.
From second release, when we as a team were ready with good understanding on the
expectations and other tech dependencies we started delivering and we could
complete 2 full flows from end to end.
We had very less amount of spillover and could complete almost everything in the
backlog, which was a big issue in earlier releases.

Anonymous · Answer

Device testing has to deal with failures at different levels: different
configurations, different provisioning, unknown initial state, devices,
different platforms, operating system versions etc.
With this many variables device testing for Google Home Devices had a high
failure rate hurting the trust in the test results.
I found that implementing extensible health checks would help ensure the tests
initiated with a healthy device with a tradeoff of adding a few more seconds
after a test. I also found that fixes to the unhealthy devices could also be
automated speeding up the device recovery and improving device availability.
I implemented this functionality adding hooks to the continuous integration
system to execute device recovery on after task executions and running device
health checks the devices were idle.
Over time the number of health checks and recovery steps increased, removing
almost completely false positives caused by unhealthy devices, recovering back
the trust on the test signals and approving them as the gating signals to
automatically release builds to canary channels.

Anonymous · Answer

At Auctane there was a decision made to migrate monolith to microservices, this
was significant undertaking for a company which was slow in technology adoption.
There was people upskilling challenges, product presure to add more features to
existing system and lack of resource in fullining the migration in a tight
deadline.
I worked an a stratergy with the product team and leadership to reduce number of
new features instead allow my team to focus on the migration part - after which
we can take the backlog items.
Once I got a greenlight, I had a solutions architect come and take 1 week course
for our team to upskill, also encouraged them to take the AWS certification in
parallel.
I had to scale up my team with external hires, and during pandemic it was
challenge to hire any resource. Once I found the right resource, I enabled them
to be productive by having a buddy dev for 2 weeks - and then learn as they go.
I trusted them to be motivated and gave them the needed tools to succeed.
There was growing pains for us as team, but in the end, we successfully migrated
to microservices on time and within budget. If given a role at Meta, I will
bring the same phiolophy of empowring my team members to do thier job and
fostering an environment of innovate.

Anonymous · Answer

The most innovative idea I've ever had was to utilize check system when
completing payment via check when starting a new policy.  When you accept a
payment that is not a secured form of payment such as credit card or western
union, depending on the state you are required to provide insurance based on the
state regulations.  For example NY requires the company to send out a
cancelation letter which gives an additional 14 days to make the payment.  
During planning season, I introduce check system based on my observation of the
system in local stores.  I was task to run a controlled pilot with the region of
240 associate for 60 days.  The results indicate immediate reduction of
cancelations due to invalid checks. Based on the success rate, the process was
implemented countrywide and the director was pleased with the results.

Anonymous · Answer

Because my current role have to monitor and track status of 8d for approx 100
cases/month, previously, we had to do manual tracking case by case and follow up
with team. My idea was we should have some tool that can assist us to monitor
cases and I asked my team to brainstorm and emphasized benefit that we will get
if we have this tool. We did small workshop and came out with idea for creating
semi-auto tracker. Then we list down task and planned it together. We had weekly
meeting to follow up pending tasks and discussed about obstacles and tried to
fix it. Until we can implement this tracker then we evaluated effectiveness of
it. We found it can reduced tracking time down 80%

Anonymous · Answer

I was designing a feature for custom creation and since it was a new space the
requirement were abstract. I was designing the experience and would always like
to prpose ideas . one of them was to elt users save their custom templates so if
there was a similar requirement that they would have it ready to go and not have
to do the process again. after discussing some ligistics with product team on
setting a limit to creaiton we imlemented it

What is the most innovative idea you have ever had? How did you come up with it and how did you go about implementing it?

Answers

Unlock Community Insights

Unlock Community Insights

Unlock Community Insights

Unlock Community Insights

Unlock Community Insights

Unlock Community Insights

Unlock Community Insights

Unlock Community Insights

Unlock Community Insights

Unlock Community Insights

Try Our AI Interviewer