Paper Slides

• Leveraged application characteristics to reduce offered load.
• Proposed and evaluated a profiling-based resource allocation mechanism for edge servers.
• Reduced offered load by 40 % and maintained application quality of service in severely loaded edge servers.

Paper Slides Code and Datasets

• Designed an edge-based architecture enabling live video analytics on small autonomous drones for search and rescue tasks.
• Proposed and evaluated four different techniques to reduce wireless bandwidth usage when offloading computation, e.g. early discard, just-in-time learning, reachback, and context-awareness.

Conference Paper Journal Paper Slides Demo Talk

• Designed and implemented a real-time IoT privacy mediation system that selectively blurs faces
• Adopted a cloudlet-based approach to achieve separation of trust and ensure platform integrity
• Leveraged object tracking to speed-up face detection and DNN-based face recognition

Paper Slides

• Measured network latency for cloudlets, small-scale datacenters located at the edge of the Internet, under Wi-Fi and 4G LTE networks
• Measured energy consumption on mobile devices when offloading heavy computation to cloudlets
• Analyzed system design trade-off between response latency and energy consumption in edge computing
• Maintained and debugged in-lab cellular base station

Paper Slides Demo

• Designed and implemented the paper keyboard UbiK, leveraging audio and motion signals received by a smartphone to detect and recognize different keys
• Performed feasibility tests and discovered multipath fading audio signatures of keystrokes on conventional surfaces
• Modified Android kernel TinyALSA audio driver to enable dual microphone recording and double feature space
• Benchmarked and evaluated UbiK, demonstrating faster typing experiences than on-screen keyboards in user studies

• Interned at Applied Privacy Research team.
• Designed an end-to-end machine learning system to automatically understand privacy-related user feedback and assign fine-grained privacy tags to users' text feedback.
• Implemented and trained a recurrent neural network (RNN) using Tensorflow to capture semantic meaning of feedback text.
• Achieved ~85% accuracy of the tag assignment.

• Developed a python framework using OpenOffice API to automatically generate SoC clock synthesis constraint files from clock signal catalogs in Excel
• Automated SoC clock cross domain checking file generation workflow

• Developed an Android application to lock all UIs in landscape mode for synchronizations of mobile phones and In-Vehicle Infotainment system
• Researched and compared different methods to achieve audio synchronization between Android phones and In-Vehicle Infotainment system

• Graduated with the Highest Distinction

• Graduated with the Highest Distinction
• Graduated with Honor in Research