Jeremy Aguilon's Page

I spent my tenure at Facebook working on Monetization Integrity. In short, we decided whether an entity on IG/FB is monetizable or not using a mix of machine learning, heuristics, and human reviewers. My projects included:

• Staging a high-profile upgrade of our core monetization policies to align with industry standards.
• Pioneering machine learning models that forecast how risky a page is to violate our monetization policies
• Working closely with peer engineers, data scientists, product managers, and policy leads to understand and close gaps in our topline metrics
• Planning, collaborating, and executing on large infrastructure overhauls of our ML-based decision serving workflows.

As an APM intern, I worked on the Assistant on automotive surfaces team. My project primarily focused on defining a new feature that will improve the daily experience of drivers. I explored use cases with other PMs, worked through design decisions with UX, and collaborated with engineering stakeholders.

At Lyft I contributed to Express Drive and Fleet Management. On the express drive end, I improved product experience through tooling for vehicle geolocation, maintenance/collision servicing, and rental reposession. Within Fleet Management, I aided the deployment and querying of Elasticsearch indices on Lyft's autonomous and rental vehicles. Holistically, I matured my understanding of datastore management, frontend design patterns, API building, and architecting brand new microservices.

Amazon Personalization houses an incredibly sophisticated product recommendation system, driven by years of A/B tests. One of the biggest challenges at Amazon scale is balancing recommendation quality with speed. I worked on incorporating more complex (and more relevant) recommendations in the parts of Amazon requiring low latency. My major deliverable was a NoSQL dataset of more than 100 million rows containing the preprocessed recommendation scores from a computationally expensive recommendation algorithm.

As a member of cloud monitoring, I improved usability of the analytics pipelines through a new pipeline monitoring service. By storing pipeline statuses in a Spanner database, I made complex tasks such as diagnosing the root cause of a cascading pipeline automatic. My entrance into this program was quite unique, as I got my interview after completing the Google Foobar challenge. Check out my portfolio for what that is!

InternBlitz is the common application for internships. I worked on the core functionality of the platform by architecting internship web scrapers, improving the application pipeline, launching a user dashboard, and maturing a messaging system prototype--all so that people could conduct an end-to-end internship search under one platform. InternBlitz was accepted into Georgia Tech's 2017 Startup Launch. I balanced my work with my internship at Amazon.

In the summer fo 2016, I worked with Vertafore on ImageRight, an insurance software that facilitates agent productivity. I helped integrate an array of cloud storage management solutions that are being implemented across all of the company's products. I also contributed to the initiative of providing high quality unit and end-to-end tests across the entire platform.