Projects
Research Projects
Binary Compression for Software Distribution
ACM SAC 2024 | Paper Link
Developed a novel binary compression technique that reduces software size by 15-30% through pattern folding. This project addresses the growing challenge of software bloat in modern applications.
Key Achievements:
- Designed pattern recognition algorithms to identify recurring patterns across multiple executables
- Implemented shared extracted pattern code mechanism allowing binaries to reuse compressed components
- Achieved compression ratios superior to traditional methods while maintaining execution performance
- Technologies: C++, LLVM, Binary Analysis, Pattern Recognition
Mobile App Size Optimization via IR Compression
IEEE RTCSA 2023 | Paper Link
Created a cross-file redundancy detection system for iOS apps using LLVM bitcode distribution format, addressing critical App Store size restrictions.
Key Achievements:
- Built shared dictionary compression framework exceeding traditional per-file compression limits
- Reduced app sizes by up to 40% for large applications
- Deployed in production environments with zero performance degradation
- Technologies: LLVM, iOS, Objective-C, Compression Algorithms
Android Memory Management Optimization
USENIX ATC 2020 | Paper Link
Designed and implemented Acclaim, an adaptive memory reclaim system that significantly improves user experience in Android devices.
Key Achievements:
- Designed Foreground Aware Eviction (FAE) to intelligently relocate memory pages
- Implemented Lightweight Prediction-based Reclaim (LWP) using machine learning
- Achieved 40% reduction in app launch latency and 60% decrease in page re-faults
- Technologies: Android Kernel, C, Machine Learning, Memory Management
Mobile Cache Management Framework
HotStorage 2020, Extended to FAST 2022 | Paper Link
Analyzed and optimized heterogeneous cache file behaviors to improve flash storage lifetime in Android devices.
Key Achievements:
- Developed online classification algorithm for hot/warm/cold cache files
- Extended to CacheSifter, reducing flash writebacks by 70%
- Maintained cache hit rates while significantly improving device longevity
- Technologies: Android, File Systems, Storage Systems, Data Analysis
Industry Projects
Microsoft Edge Enterprise Mobile
Microsoft | 2022 - 2024
Developed enterprise mobility features for Microsoft Edge on Android and iOS platforms.
Key Achievements:
- Implemented security protocols improving corporate data protection by 40%
- Built cross-platform synchronization for seamless desktop-mobile workflows
- Shipped features adopted by 200+ enterprise clients within first quarter
- Technologies: C++, Chromium, Objective-C, Java, Android/iOS
iOS Applications Portfolio
Independent Developer | 2025 - Present
Developed and published multiple iOS applications on the App Store.
monait
Personal finance management app with AI-powered insights
- Intelligent expense categorization and budget tracking
- AI-driven financial insights and recommendations
- Clean, intuitive UI with dark mode support
- Technologies: Swift, SwiftUI, Core Data, CloudKit
DoggyDaily
Pet care tracking application with health monitoring features
- Comprehensive pet health and activity tracking
- Medication reminders and vet appointment scheduling
- Photo journal with milestone tracking
- Technologies: Swift, HealthKit Integration, Push Notifications
Omacase
E-commerce platform for custom phone case design
- Real-time design preview with AR visualization
- Integration with payment gateways and shipping APIs
- Custom image upload and editing capabilities
- Technologies: Swift, ARKit, Stripe API, Firebase
Open Source Contributions
- vLLM: Active contributor to the popular LLM serving framework
- Chromium: Contributed patches for mobile browser optimizations
- LLVM: Submitted improvements to bitcode compression modules
Tools & Utilities
Memory Profiler for Android
A lightweight tool for profiling memory usage patterns in Android applications, helping developers identify and fix memory leaks.
Binary Pattern Analyzer
An analysis tool that identifies recurring patterns in compiled binaries, useful for compression and optimization research.