Research Projects
My research sits at the intersection of Machine Learning and Electronic Design Automation (EDA), with a particular focus on hardware security. I explore how AI can revolutionize traditional EDA workflows while ensuring security and trustworthiness in modern computing systems.
ML for Design Automation
Next-Generation EDA Algorithms
Core Focus: Developing ML-powered algorithms for complex EDA challenges
Current Work: Designing novel floorplan algorithms using machine learning techniques for placement and routing optimization
Research Directions:
- Floorplan Optimization: ML-driven placement and routing algorithms
- Design Space Exploration: AI-driven optimization for multi-objective design
- Formal Verification: ML-enhanced verification and refinement relation discovery
Innovation: RE3 algorithm for automatic refinement relation discovery (DAC 2025)
ML-Driven Hardware Protection
AI-Enhanced Security & Design
Core Focus: Leveraging machine learning to advance hardware security and applying hardware security techniques to protect AI systems
Key Innovations:
- eFPGA Redaction: Graph neural networks for physical-aware IP protection
- Hardware-to-AI Security Transfer: Extending traditional circuit protection techniques to secure LLM parameters and generative models
- Intelligent Design Flows: ML-guided security-aware design methodologies
Current Work: eFPGA-based hardware IP protection techniques, bidirectional security applications between hardware and AI systems
Impact: Creating synergy between traditional EDA security and modern AI protection
Hardware Security Foundations
Logic Locking & Circuit Protection
Core Focus: Developing robust logic locking schemes and attack methodologies for IP protection
Key Contributions:
- ObfusLock: First framework achieving simultaneous security, obfuscation, and efficiency
- DE2: Novel SAT-based decryption using functional specifications
- DNN Logic Locking: Security analysis of neural network protection schemes
Publications: DATE 2023, DAC 2024, DATE 2025
Status: Ongoing research with 3 published papers
Cross-Domain Applications
AI for Healthcare
Core Focus: Applying ML expertise to solve real-world problems beyond hardware
Achievements:
- Multimodal Medical Diagnosis: Bayesian networks + deep learning for skin disease detection
- 19.3% accuracy improvement over pure deep learning approaches
- Extensible architecture supporting diverse clinical information integration
Publications: ISBI 2023, BIBM 2024
Impact: Demonstrating transferable ML expertise across domains
Research Vision
I believe Machine Learning and EDA are mutually empowering. ML can enhance traditional design automation, while hardware techniques like security methods can strengthen AI systems. My goal is to bridge these domains and create practical solutions that industry can adopt.
Key Insights:
- Bidirectional Innovation: ML enhances EDA efficiency; hardware security protects AI models
- Industry Relevance: Research should address real-world challenges with deployable solutions
- Practical Impact: From theoretical frameworks to production-ready tools
My recent completion of eFPGA-based hardware security research exemplifies this industry-focused approach, developing techniques that can be integrated into existing design workflows.
This research is conducted at the NuLogiCS Research Group, Northwestern University, in collaboration with industry partners and domain experts.