As a Technical Program Manager at Peloton:

• I led all activities for the launch of the Pause feature on bikes and treadmills (the most user requested feature ever)

• I managed was the computer vision powered Rep Counting feature. I managed the technical implementation of this feature, from model architecture to data collection and training, and coordinated across our marketing and UI/UX teams to launch the product. After its release in September 2022, I continued to optimize model capabilities.

For my senior thesis at Scripps College:

• I investigated whether we could computationally derive a preferred factorization of the quantum mechanical Hilbert Space into subparts.

• I computationally implemented a streamlined version of the “in principle” proposal from the paper Quantum Mereology: Factorizing Hilbert Space into Subsystems with Quasi-Classical Dynamics. I coded a platform to assign a score to the tensor product structure (TPS) of a state-space based on how robust of a pointer observable it admits. I extended the paper by applying a basin-hopping algorithm to optimize over the space of all TPSs. 

As a research intern working primarily with the Volo Foundation:

•  I was tasked with the open ended goal of understanding if ragweed pollen was increasing due to climate change and resulting in higher rates of asthma. 

• Through my work collaborating with experts in the field, I obtained and analyzed existing datasets.

•  I realized there was a dire lack of ragweed data collection, leading to our decision to invest in indexing sites. 

As a software engineering intern as Google:

• I designed, implemented, and tested a Python library that simplifies the process of creating workflows for orchestrating moderation of user-generated geographical data.

• In addition to my full time project,  I decided to do a 20% project with the Artists + Machine Intelligence Team, curating and presenting a survey of artists who use Google Maps in their creative work.