Could you tell me about a project you are very proud of?

In my role as Team Lead Manager at Flexport, I led a team of six engineers tasked with building a Kafka-based streaming data infrastructure deployed in Kubernetes. We encountered a significant challenge at the start of the project: integrating with existing products while meeting the requirements for products still in the design and planning phases.

To address this, I implemented several key strategies:

Clear Communication: I set up design review sessions with the entire engineering organization to ensure everyone was aligned and any potential issues were addressed early.

Team Collaboration: I encouraged regular pair-programming sessions and implemented pairing rotations to enhance collaboration and knowledge sharing within the team.

Agile Approach: We adopted an Agile methodology, breaking down the project into manageable sprints. This approach helped us tackle the problem incrementally and adjust our plans as needed.

Early Data Contract Discussions: We initiated discussions about data contracts with other teams early in the process. By co-publishing data contract proposals and seeking feedback, we ensured that our solutions were well-integrated and met cross-team requirements.

Task Management: I made sure tasks were broken down into manageable sizes with clear descriptions and acceptance criteria. This approach helped maintain team velocity and clarity on deliverables.

Transparency and Accountability: We conducted bi-weekly public demos to track our progress and maintain transparency. These demos allowed us to gather early feedback and hold ourselves accountable to our target dates.

By employing these strategies, we successfully launched our platform and onboarded the first production workflow within three months. The production workflow was to provide reporting for our customers. We reduced reporting latency from 24 hours to near-real-time. Additionally, we introduced a self-service template and comprehensive documentation, which reduced the workflow onboarding time from one week to less than two days.

This project was a team effort, and the team was highly motivated and proud of the result. It taught me the importance of breaking down large projects into smaller, manageable pieces, which has since improved my approach to leading technical teams through complex problems. Regular demos and project updates were crucial to the success of this project, as they allowed us to gather feedback early and stay on track with our deadlines.

Situation: The North America Transportation (NATS) organization, which is dedicated to managing the design, speed, and connectivity of the transportation network, previously did not have a formal process for reviewing and tracking the impact of new pilots and initiatives. This led to issues with communication, lack of network-wide visibility, and unknown impacts on key metrics.

Task: I was tasked with creating a new pilot tracking program from scratch to address these challenges and improve the NATS organization's ability to evaluate and scale successful initiatives.

Action: I built a comprehensive pilot program within the Asana project management software, consisting of four main components:

1. Pilot intake with a white paper outlining success criteria and risks

2. A monthly pilot review panel with NATS leadership

3. Ongoing pilot tracking for approved initiatives

4. Post-mortem and scaling process for successful pilots

Result: The new pilot program was launched in Q3 2022 and has already demonstrated a 35% cost savings in transportation execution costs through the successful implementation of new innovation programs. The structured review, tracking, and scaling process has improved visibility, risk assessment, and performance monitoring across the NATS organization.

The accomplishment that I am most proud of is when my manager asked me to move to Malaysia to set up a new Factory and the Operations department in Kuala Lumpur.

It was my job to get all licenses and permits required to allow our factory to run under the country's regulations. I had to define the different roles needed like Product, Process, Quality, and Planning engineers. In addition, it was also my responsibility to manage the transfer of 3 different product lines from China to Malaysia.

To make that happen, I kicked off the project with the stakeholders to define the goals, budget, and plan for the opening of the new factory. I defined all milestones, and I collaborated with HR team to define the total HC needed, the different roles, the timeline for each incorporation, and the budget. With the Legal department, I worked closely to get the licenses and permits required for our company to operate in Malaysia. Together with the Operations team, I worked on the factory layout, material readiness, staff training, and the assembly lines transfer plan from China to Malaysia.

After 9 months of deep collaboration with cross-functional teams, I completed the project on time and within budget. It was a great learning opportunity that challenged my program management and leadership skills. Finally, the transfer resulted in a cost down of 5% for our customer in the products transferred to Malaysia.

I want to talk about a project to scale our push notifications backend platform.

One day, we were getting paged for replication lags. This was a sign for database clusters getting pushed to their limit. We double-checked whether this was a one-off or we are facing a scaling issue. We concluded we were facing an issue. The problem here was not only we needed to scale up but we sharded our database just about a year ago with what we thought would be fine for at least 3 years. Instead of just working on scaling the database, we talked to our clients and saw how much more volume they were demanding and it came to our surprise our clients were actually looking for 3x the current quota. Instead of just scaling our database, we decided to look for a longer-term solution.

My role in this project was not only accountable for delivering the project but also making sure technically the new solution made sense and was not over-engineered. I was more technically involved than in other projects and asked a lot of challenging questions while empowering ICs to look for creative solutions involving re-architecture. I had my team investigate options while I was talking to my manager about this work. Our clients were asking for 3x the traffic so that was our goal for this project. We also set a stretch goal of supporting 5x the traffic if possible. After investigations and talking to experts, we decided to utilize an auto-scalable cloud component to replace some of our in-house managed components such as the database and queuing infra.

The project involved some architecture to utilize cloud components. Especially, our data access pattern made it possible to efficiently utilize DynamoDb with its autoscalability allowing us to actually meet our stretch goal. We faced some challenges around hidden limitations that were not documented well but we were able to overcome those by talking to experts and implementing best practices.

After delivering the project, we talked to our clients again and allocated 3x quota to teams that were demanding it. One team had a campaign that took 10+ hours to send out but is now sending the whole campaign within 1 hour increasing user engagement and being able to better target time window to send messages.

We learned from this project that while the cloud provides very interesting tools and solutions, they don’t magically solve scaling problems. Our experience and learnings from this project helped us tremendously with our future cloud migrations.

While I was working on the Fuchsia Team I had to architect and implement test beds, services and workflows to simulate the upgrade process of the operating system from Linux to Fuchsia. The manual validation process required a sequence of button presses to trigger the upgrade process and downloading and installing the production version of the operating system.

I had to collaborate with the team building the upgrade engine to allow replacing the upgrade APIs to a testing version. I had to design a new version of the upgrade APIs to ensure we could validate downgrading, upgrading and controlling the versions for testing.

I also had to collaborate with the electronic engineers of the team to modify the smart display devices to programmatically generate the button push events.

This was a very complex workflow that was going to be used to validate the upgrade process for devices already in production around the world and even small failures could render the devices unusable.

The test beds, the test upgrade service and multiple rounds of validations and fixes made the upgrade process a success with more than 99.5% success rate. In fact it was such a success that nobody noticed it until people found in the news their devices were already running Fuchsia. I’m really proud that the level of detail and the hard work replicating the upgrade progress made the infield migration a success.

A project I am very proud of is creating and implenting the perferred employer program at my community. In part of creating this program I had to dive deep into our communites demographcs as well as coming up with a list of employers that have employees we would like to attract. I focused on large surrounding businesses like state and medical employees. As part of the PEP program I offered free application fees as well as a $400 move-in credit. Once I saw what an impact this made on occupancy. I increased the amount of the resident referral program to get the employees we attracted through the PEP to start referring other emplyees to us. The resident referal program went from a flat $500 referral fee to a tiered program that started at $500, went to $750, then to a $1000 and for the fourth referral $1500. This increased revenue by around $200,000/year which is nearly 20%

I was involved in a project where we have build a statistical model to leverage the historical data related to meters that print labels and predict the likelihood of the different issues in the meter so a correct kit can be ordered to in first attempt to service person visit

I am particularly proud of the last project achievement. The purpose was to integrate a translation management tool to handle outsourced translation request. The tool was very complex and we had to map its items with our datas and coect the tool with all our applications and integrate all our processes like translation requests, reinternalization, cacelation, assigments, creation of purchase orders and invoices....

The project is a total success and is actually used by our stakeholders.