Lekha Patel

Lekha Patel

Statistician

Sandia National Laboratories

Biography

I am a senior statistician at Sandia National Laboratories. My research interests broadly lie in the fields of computational statistics, simulation-based statistical learning (particle methods, Monte Carlo approaches, Approximate Bayesian Computation) of stochastic processes and Bayesian nonparametrics. Other areas of interest include temporal-spatial statistics, extreme value analysis and Bayesian variable selection. Recently, I have developed an interest in physics-informed statistical learning. Applications of my work include in bioinformatics and biological imaging, cyber-security, climatology and the social sciences.

Download my resumé.

Interests
  • Computational statistics
  • Simulation based statistical learning
  • Bayesian nonparametrics
  • Stochastic processes
  • Temporal-spatial statistics
Education
  • PhD in Statistics, 2019

    Imperial College London

  • MSci in Mathematics, 2015

    Imperial College London

Experience

 
 
 
 
 
Statistical and Machine Learning (ML) consultant
Jan 2022 – Present Boston, MA
Devising statistical and ML algorithms for DNA sequencing, early disease detection and health monitoring.
 
 
 
 
 
Statistician
Jan 2020 – Present Albuquerque, NM
Devising novel statistical methodologies and inference mechansisms for applications in national security.
 
 
 
 
 
Machine Learning (ML) consultant
Aug 2019 – Jan 2020 London, UK
Implemented automatic skills labeling scheme with sparse and inbalanced data.

Recent Publications

Quickly discover relevant content by filtering publications.
(2022). Development and characterization of a tabletop fog chamber at Sandia National Laboratories. Proc SPIE.

DOI

(2022). Parameter Estimation of Binned Hawkes Processes. Journal of Computational and Graphical Statistics.

Cite DOI

(2021). Spatio-temporal extreme event modeling of terror insurgencies. Preprint.

PDF DOI

(2021). Blinking statistics and molecular counting in direct stochastic reconstruction microscopy (dSTORM). Bioinformatics.

PDF Cite DOI

(2020). Assessing Extreme Value Analysis to predict rare events from the Global Terrorism Database. Proc. JSM.

PDF Cite

(2019). A hidden Markov model approach to characterizing the photo-switching behavior of fluorophores. Annals of Applied Statistics.

Cite DOI

(2018). Bayesian filtering for spatial estimation of photo-switching fluorophores imaged in Super-resolution fluorescence microscopy. In 2018 52nd Asilomar Conference on Signals, Systems, and Computers.

PDF Cite DOI

Contact