Cashtags: Designed and implemented Android-based mobile security system to provide secure input and prevent sensitive data leaks.
- Modified system display rendering data path to support widget view interception.
- Demonstrated practical system data storage and overhead requirements.
- Utilized novel crowd‐sourced product testing paradigm to enhance product.
DexDiff: Co‐designed malware detection system to assess security impacts of Android vendor customizations.
- Prototyped system to structurally compare Android binaries and highlight control flow graph changes.
- Analyzed sixteen pairs of Android binaries from stock Android against closed‐source 3rd party Android devices.
- Evaluated system prototype for detection accuracy, scalability, and efficiency.
Mobile Privacy Survey: Designed and implemented 600+ participant human subjects survey on mobile device usage and privacy practices.
- Developed Android and web questionnaire and scalable server back‐end to collect participant responses.
- Acquired Human Subjects Committee (IRB) and National Science Foundation certifications and authorizations.
- Performed rigorous statistical analysis and significance testing on multidimensional dataset using custom tools.
Long Term Mobile Usage Study: Designed and executed three month human subjects tablet usage and privacy practices monitoring system.
- Created and extensively tested custom Android Nexus tablet firmware to transparently collect usage data.
- Developed scalable, encrypted, and anonymized server back‐end for data reporting.
- Implemented system hooks for remote administration and analyzed and debugged device problems.
ContextProvider/Bio-Environment Android Tracking: Designed and developed context‐aware health monitoring and emergency alert mobile framework.
- Integrated hardware sensors, web services, and biological data sensors to accurately detect context.
- Developed multi‐level alert system to elevate system response based on severity of detected anomalies.
- Analyzed system accuracy using human subjects and evaluated system power and storage overhead.
TrueErase: Collaborated on development and testing of full‐storage‐data‐path, per‐file secure deletion system.
- Prototyped legacy‐compatible framework that propagates file‐level information to the block‐level storage layer.
- Verified implementation correctness and evaluated system I/O bandwidth overhead.
Multi-core I/O Analysis: Designed and developed experiments to isolate the cause of observed multi‐core I/O bottleneck.
- Created experiments to isolate the cause of multi-core I/O bottleneck in benchmark testing.
- Modified existing benchmark utilities to support and test process/thread core affinities.
- Gained proficiency in system and kernel profiling, performance evaluation, and statistical analysis.
MusicReco: Co‐designed and developed music analysis, classification, and recommendation system
- Developed recommendation system incorporating audio processing with efficient preference prediction.
- Utilized machine learning tools to perform automatic genre classification for audio files
- Incorporated fast k-nearest neighbor calculation and majority voting to determine user likes/dislikes
Mitchtech, LLC: Launched, my own software development and consulting firm.
- Developed over a dozen Android applications with over 500,000 market downloads.
- Lead team of Android and Linux developers and collaborated on multiple software and hardware start‐up ventures.
- Authored over 100 technical articles and tutorials, many featured in popular technical periodicals.
- Developed and published over 50 open source software development projects.