Impact¶
Driving Measurable Outcomes Across Research and Industry¶
My work is measured not by the models I build, but by the real-world outcomes they enable. Across supply chain optimization, sustainability analysis, and AI-driven decision systems, the through-line is impact: quantifiable, actionable, and enduring.
Key Impact Areas¶
Supply Chain Resilience¶
Problem: Traditional binary disruption models miss the majority of real-world supply chain risks — partial capacity losses that silently erode performance.
Action: Developed partial disruption modeling using stochastic mixed-integer programming to capture the full spectrum of supply chain vulnerabilities.
Result:
- Identified previously invisible risk exposure in pharmaceutical supply chains
- Enabled more cost-effective mitigation investments through targeted strategies
- Provided a realistic risk landscape for strategic planning and capital allocation
Cost Reduction & Efficiency¶
Problem: Large-scale manufacturing at Dow required continuous process optimization to maintain competitiveness under shifting market and regulatory conditions.
Action: Applied advanced optimization and data analytics (Python, Pyomo, SQL) to real-world industrial processes, bridging R&D insights with plant-floor operations.
Result:
- Achieved measurable process efficiency improvements in target operations
- Bridged the gap between R&D and plant-floor implementation
- Demonstrated scalable optimization deployment in industrial settings
Sustainability Improvements¶
Problem: Industrial operators needed to quantify the true cost-sustainability trade-off to make informed capital allocation decisions and meet regulatory requirements.
Action: Built integrated LCA/TEA models with AWARE water footprint methodology, mapping cost-sustainability frontiers across process alternatives.
Result:
- Mapped cost-sustainability trade-off frontiers for multiple process configurations
- Quantified water footprint using the AWARE methodology
- Delivered actionable recommendations adopted by policymakers and stakeholders
Institutional Decision-Making¶
Problem: Graduate advisor-student matching relied on ad hoc methods, leading to misalignment, attrition, and suboptimal research outcomes.
Action: Designed an agent-based simulation framework with autonomous agents representing students and advisors with heterogeneous preferences and constraints.
Result:
- Improved match quality over baseline preference-based methods
- Delivered a scalable framework applicable to other allocation contexts
- Highlighted multi-agent modeling as a tool for complex institutional decisions
Impact Snapshot¶
| Project | Action | Outcome |
|---|---|---|
| Pharmaceutical Supply Chain | Stochastic partial disruption modeling | Revealed hidden risks, improved resilience planning |
| Dow Process Optimization | Advanced optimization & data analytics | Measurable efficiency gains in manufacturing |
| LCA/TEA Sustainability Analysis | Integrated environmental-economic modeling | Data-driven sustainability recommendations adopted |
| Advisor-Student Matching | Agent-based simulation framework | Improved match quality, scalable framework delivered |
Philosophy¶
Impact Principle
Every model, every analysis, and every framework I build is guided by a single question: How does this create real-world value?