University of California, Berkeley

Beverly Hills High School

Software Engineer @ Hex

AI And Product

Software Engineer

Nines is applying world-class ML and AI expertise to the field of medicine via teleradiology. Backed by Accel Partners and 8VC and partnerships with several of the top health systems in the entire US, we are bringing the power of modern AI to clinical settings with industry-leading results.

As a software engineer I worked full stack on our entire pipeline with a focus on backend systems and customer integration endpoints. 
This included projects such as 
- Designing and managing our interface for ingesting millions of medical images a day 
- Internal tooling for managing and viewing clinical data 
- End to end product features that augmented the radiologists reading experience increasing efficiency and ease of use 
- Designing and implementing customer interfaces for ingestion of data from customer specific formats to standardized internal formats 

Extensively used: Java, Python, Typescript, Kubernetes, Protobuf, gRPC, Postgres

Nines

Amplify Partners is an early-stage venture capital firm that works side by side with first-time founders and serial entrepreneurs to transform their groundbreaking ideas into companies that win. Over thirty-five founding teams in infrastructure and data-driven technology have selected Amplify as their partner on the path to success. With expertise gained from decades of experience, the Amplify team provides critical help with engaging early customers, establishing product-market fit and attracting the right talent to build companies that win.

Learn more: http://www.amplifypartners.com 

Assisted in technical due diligence, participated in investment team meetings and contributed to deal discussion. Focus in machine learning applications.

Amplify Partners

University Associate 

-Expanded MCMC inference capabilities for probabilistic programming language library by adding support for chain inference, implemented parallel tempering, research and added convergence diagnostics, and added multi-model distribution based models for benchmarking.

Facebook

Probablistic Programming Intern

-$70 million backed Series B autonomous drone company.   
- Implemented data generation pipeline used to create 50k synthetic stereo image dataset using UnrealEngine. 
-Created benchmarking tools to aid in state of the art deep stereo model testing and fine-tuning.

Skydio

Computer Vision Intern

Selected as one of 25 Accel Scholars in my cohort.

Accel

Accel Scholar

-      Intern within Computer Vision working closely between Mobile Vision and Comuter Vision Platform teams 
-	Expanded internal infrastructure and online interface within FB’s internal computer vision platform to support mobile vision model training and direct deployment on a mobile phone with push of a button using transfer learning with deep convolutional nets and logistic regression. 
-       Full stack project included using Python, C++, Objective-C, PHP, React, SQL, and GraphQL for backend, frontend and mobile development changes and Caffe2 for creating and training models

Olivia Koshy