Prashant Arunapuram Email & Phone Number

Mountain View, California, United States

Newark, California, United States

• Proposed and delivered multi-task framework for performing occlusion detection and lane line detection simultaneously, eliminating the need for a second model ----> Shipped 3 models for the first Lane Change Assist production release, for rear camera sensor degradation and rear camera online calibration• Technical Lead for new pedestrian detection model on surround view fisheye cameras, directing a team of 2 engineers

Newark, California, United States

• Developed and delivered 4 models that determine the occlusion status of front and side ADAS cameras on the Lucid Air‎‎----> Owned pipeline from creating annotation guidelines and running model training, to interprocess integration, deployment, and optimization----> Interfaced with product team about requirements to understand customer needs and improve performance----> Designed models for NVIDIA safety mode, and features to ASIL D standards• Worked cross-functionally with other perception sub-teams, sensor fusion team, and maps team to coordinate behavior based on results of each inference• Supported development of general perception stack, technical debt cleanup, visualizer improvements, refactors, and managing system level messaging

Pittsburgh, Pa

• Led development of computer vision algorithms for a mobile robot arm system; used in applications of aircraft water panel, construction vehicle cockpit, and helicopter fuel port• For each application, detected components using classical computer vision, translated environment from 2D images to 3D space using stereo geometry, did trajectory planning, and completed physical manipulations• Performed demos for contractors with fantastic feedback----> Demonstrated manipulation and vision capability of 6DOF robotic arms by performing tasks such as connecting a hose to a fuel port, refueling a water tank, and operating a large front loader.

Pittsburgh, Pa

Researched published articles on current state of the art methods in multi-view stereo for efficient large scale reconstructions of 3D images.

Pittsburgh, Pa

Wilmington, Delaware

- Researched and recommended enterprise Virtual Personal Assistant tools, similar to Apple’s Siri and Amazon’s Echo, for enabling knowledge sharing across the organization using linked data- Used Neo4J graph database with Cypher Graph Query Language (similar to SQL).

Pearl River, New York

- Developed a web application for oncology research scientists to query, annotate and visualize the antibody sequences.- Used HTML, PHP, SQL and JavaScript to interact with ANARCI open source to obtain annotated sequences and other attributes for a given antibody sequence- Created a graphical view NextProt Feature Viewer to show features and frequencies for antibody sequences

Master Of Engineering - Meng, Computer Engineering

Activities and Societies: CMU Raasta• Studied Computer Engineering, Completed 5 year Masters Degree • Concentrations in Machine Learning.

B.S., Electrical And Computer Engineering

Projects: Web Proxy Server – Implemented a fast web proxy server to mediate requests and responses between clients and servers utilizing.

High School

Member of several teams during my high school years at Henderson. - Varsity team member of the academic team. We came in 2nd place.