I'm currently an Associate in the Buildings & Structures practice at Exponent, a leading scientific and engineering consultancy. You can find my full professional bio here.
You can find my CV here and my GitHub here.
Connect with me at linkedin.com/in/gbhattacharjee.
Contact me at gbhattacharjee(at)exponent.com.
Research
Keywords: risk analysis, probabilistic risk assessment, uncertainty quantification, decision-making under uncertainty
I specialize in developing probabilistic risk assessment methods and mitigation approaches for complex regional infrastructure networks or other systems subject to uncertain hazards. I’ve worked on seismic, wind, environmental contamination, lightning, and multi-hazard analyses for roads, electric utilities, and underground pipelines. I typically work on risk models from conceptualization through implementation (typically in Python).
My expertise includes disaster risk management, uncertainty quantification, statistical analysis, machine learning, and qualitative research methods. I’ve also worked on projects assessing the cause of damage or collapse of structures following flooding, corrosion and deterioration, and impact.
As a graduate researcher, I developed network performance-based methods for assessing and mitigating the seismic risk of regional road networks. My work in that area included using global sensitivity analysis to prioritize bridge retrofits and developing surrogate models to reduce the computational burden of traffic simulation in the context of probabilistic risk assessment. I also contributed to refinements in the metrics with which post-earthquake road network disruptions are characterized.
Personal
In my free time, I read a lot and write a little less. I love making prints and conversation. You can find me walking all around New York in any season except August, when I decamp from the city in search of air conditioning. My life goals include penning an advice column and taking a decent selfie.
I write a very occasional newsletter on what I’m reading at Substack.