Software engineer on the Platform team at LMR Partners, a global macro hedge fund. I work across the stack to improve the performance, reliability, and operational stability of internal applications in a demanding, latency-sensitive environment. My work spans proactive engineering — identifying and resolving bottlenecks before they become incidents — as well as owning ad hoc operational issues end to end. I bring a breadth-first approach to platform engineering, taking on whatever the system needs to keep things running at the standard a top-tier fund demands.

Leading the Ads Engineering team within the Bloomberg Media organization.

Technical Environment: Java, Kotlin (SpringBoot), NextJS, Typescript, React, React Native,

AWS (EKS, ElastiCache), GCP (BigQuery, GCS), Rest API

● Built up the Ad Engineering team handling ads across all Bloomberg Media platforms including

Web, Mobile Apps, OTT Apps, SmartTV Fast Channels, CarPlay, podcasts, etc.

● Designed, built and managed Kotlin based asynchronous backend API services handling 50M

ad requests per minute.

● While the role is a managerial one, spent 50% of time writing code and mentoring other engineers

to train them in multi-threaded programming (using Kotlin Coroutine and transitioning into Virtual

Threads in JDK21)

● Services are architected with stability, resiliency and performance as the primary drivers -

Currently the services are running at 99.999% uptime with explicit failover (hot-cold set up)

across 2 different AWS availability zones.

● Proposed and productionized centralized architecture where ad targeting is performed across all

platforms using auto-scaling Docker containers on a Kubernetes cluster (EKS)

● Designed an internal event tracking system where user activities as well as ad metrics are gathered

across all platforms, persisted in BigQuery and visualized using Mode UI.

● Daily interaction with the Commercial and Product teams to establish modernized ads architecture

with new ad products to become the top technology firm among the media publishers.

● Currently exploring AI models to determine content types of different media assets

Technical Environment: Java (SpringBoot), Kotlin, React, AWS, gRPC, ECS

● Led multiple teams dealing with creating and selecting sponsored ads on Amazon's retail sites with

average annual revenue of $30B.

● Stayed hands-on technically by proposing architectural changes to the workflow framework,

testing gRPC endpoints, designing and creating the data warehouses, and reviewing codes.

● Managed standalone ads auction service that handles over 100K transactions per second and

provides final decision on which ads are displayed on Amazon.com in real time.

● Constantly engaged in enhancing the infrastructure to increase scalability and resilience, i.e

transitioning from NLB based ECS cluster to Envoy coordinated logical container clusters.

● Helped convert existing legacy Java projects to Kotlin.

● Engaged in data analysis of Amazon’s ads activities such as shopper behaviors around ads clicks

and conversions. These analyses drove experiments to change how ads are placed in different

places of the pages by using better relevancy scores per ad or predictions based on different

models.

Technical Environment: Scala, Python, GCP, BigQuery, gRPC

● Managed two teams with entirely different products and focuses

● In Jan 2020, created the first team from scratch and as of Apr 2021, the team is at 9 people.

● In Feb 2021, created a second team. This team is projected to grow to 6 people by EOY.

● Proposed, designed and implemented the Financial Data Lake with modern data engineering

pipelines and BigQuery tech stack.

● Assumed the role of both the EM and PM initially then hired a dedicated PM once the team was

stable and mature.

● Proposed and implemented new workflows within the FP&A team including FX rates

calculations, forecasts sharing, quarterly metrics, and more. The result was the enablement of

many business use cases that were not possible previously.

● Rebuilt the data flow from Kyriba to the Treasury dashboard. This reduced the data error rate

from 2~3 per month to nearly 0. This was critical because those errors were often occurring during

the month-end processes and forced the business to scramble to finish in time.

● Successfully replaced the legacy instance of MS SQL server that more than 10 business and

engineering teams were relying on. The users are now fully on the new Financial Data Lake. This

not only reduced the cost by 85%, but it also increased productivity significantly due to the

improvements in speed, accuracy and reliability of the data.

● Created the fully customizable ingestion pipelines from Adaptive, Coupa and Netsuite to

BigQuery. They were built in Scala using the Zio framework.

● Rebuilt the Functional P&L report as well as the Balance sheet report for the Accounting team in

Looker.

● Built metadata ingestion pipelines for podcasts

• Proposed and implemented the company's data strategy.

• Created the data warehouse and established ETL pipelines for most of data that exists in the firm.

• Isolated all of data related infrastructure in a separate AWS account.

• Deployed a Spark cluster for training machine learning models.

• Established a data access protocol and means to query data per each team.

• Managed a team of data engineers and data scientists.