Owner for core financial metrics/KPI engine, powering VCs to track portfolio performance. Improved validation with database triggers, developed interfaces for writing/retrieving metrics across the app, and architected data pipelines for sharing and syncing millions of metrics

Owned internal infrastructure and tooling, building multiple data importers that our services team relied on for customer implementations, libraries for Postgres data anonymization and S3 file exports, and more

Worked on refactor of our cap table management product, boosting average P50 and P99 endpoint performance by over 60% while simultaneously implementing a library for safely double writing across multiple DRF API endpoints

Rolled out static type-checking across to over 20% of our Python codebase with Pyright, reducing typing-related errors by over 40%, while automating the process of annotating 100% of our legacy code by generating types from our test suite

Utilized Python/Django with OpenPyXL to add full currency conversion support in our platform’s Excel exports for customer’s financial metrics and investment details along with adding settings to configure this using React

Implemented granular edit and view permissions system across all of our application’s pages and APIs utilizing Python/Django, React, and TypeScript along with adding comprehensive frontend and backend unit testing coverage

Built out new zip export system that enables downloading files from S3 while allowing for arbitrarily large exports and avoiding persisted file storage utilizing Python, TypeScript, Boto3, and Celery queues

Build out support in the Logs testing framework to support multiple CPU architectures

Increased C++ Logs Analysis API test coverage (unit and integration tests) for rollout of new features

Designed dashboards with Google Data Studio using SQL to visualize and track metrics that show the growth of new feature usage and Logs efficiency gains

Utilized frameworks and libraries such as gRPC, Protocol Buffers, GoogleTest, and Abseil

Worked on analytics and prediction models for the engagement of deals served to users

Utilized Python with pandas, NumPy, SciPy, and Matplotlib to clean and analyze a dataset with 100k+ entries

Designed, implemented, and trained a machine learning model using scikit-learn to predict whether the click through rate of a given Google Search/Maps deal will be high or low for the purpose of ranking

Navigated the process of model research and feature engineering, starting with a machine learning model with 1% accuracy and reaching a model with 94% accuracy and a 75% recall score