Walter Landry Ph.D.
California Institute of Technology
626-376-7498
wlandry@caltech.edu
https://wlandry.net

Experience Summary
------------------
Developed high performance programs on a wide variety of serial and
parallel architectures. Extensive experience with C++ and numerical
methods. Moderate experience with parsers, multithreaded programming,
Python, Docker, high performance web services, SQL, databases, and
Google Cloud.  Lead a technical team to enable public international
access to multi-petabyte astronomical databases, updated
daily. Publications in C++ programming, astronomy, geology, and
physics.

Selected Public Projects
------------------------
SDPB: A highly parallel, arbitrary precision semidefinite solver in
      C++ designed for problems in Quantum Field
      Theory. https://github.com/davidsd/sdpb

Gamra: A highly parallel, finite difference, adaptive mesh refinement
       C++ code for modeling deformation during an
       earthquake. https://www.wlandry.net/Projects/Gamra

Gale: A parallel, finite element C++ code for modeling mountain
      building, rifting, and subduction.
      https://geodynamics.org/cig/software/gale

blocks_3d: A heavily optimized, multithreaded C++ code to generate
           conformal blocks for SDPB.
           https://gitlab.com/bootstrapcollaboration/blocks_3d

libADQL: A C++ library to parse ADQL (Astronomical Data Query
         Language), a variant of SQL 92 with extensions for geometric queries
         on the sky.
         https://github.com/Caltech-IPAC/libadql

SuGrSonic: A highly optimized C++ library that uses Intel intrinsics
           (SSE2, SSE3, SSSE3) to model sound on a grid, requiring only 5 cycles
           per grid point per iteration.
           https://gitlab.com/wlandry/sugrsonic

libhires: A C++ image processing library for converting observations
          from the Planck spacecraft into high resolution astronomical
          images. Libhires leverages mlpack and a variety of statistical methods
          (simple binning, Elastic Net, Maximum Correlation Method) to
          reconstruct images with resolution better than the underlying
          instruments.
          https://github.com/Caltech-IPAC/libhires

libtinyhtm: A library for efficient storage and lookup of datasets
            with billions of stars.
            https://github.com/Caltech-IPAC/libtinyhtm

Education
---------
Ph.D. in Astronomy, Cornell University, Ithaca, NY
B.S. in Astronomy, California Institute of Technology, Pasadena, CA

Select Work Experience
---------------
2018-Date: Research Software Engineer, Theoretical Physics, Caltech

• Improved the scaling and performance of SDPB, a semidefinite solver
  for Quantum Field Theory, scaling it to hundreds of cores and
  enabling new classes of problems to be solved. This also involved
  modifying Elemental, a massively parallel linear algebra library, to
  fix bugs and use faster arbitrary precision arithmetic libraries.

• Ported SDPB to Google Cloud.
  https://cloud.google.com/blog/topics/hpc/introducing-hpc-vm-images

• Converted Mathematica codes scalar_blocks and blocks_3d to highly
  multithreaded C++, reducing runtime from a year to days.

• Installed and maintained SDPB on more than a dozen HPC systems
  throughout the world.

• Updated Python scripts for new versions of SDPB and Python.

• Provided support to scientists across the world, including providing
  personal assistance, creating Docker images, and giving
  presentations on progress and new features at conferences.


2012-2018: Lead Software Engineer, NASA/IPAC Infrared Science Archive, Caltech

• Managed a team of back end software engineers for IRSA, the NASA
  funded organization tasked with archiving and publishing NASA data
  in the infrared. This included:

  – Ingesting multi-petabyte data from multiple sources, with daily
    updates.
  
  – Maintaining, updating, and modernizing the publicly visible REST
    back end service (up to a million hits per day) that also powers
    the web front end.

  – Interfacing with staff astronomers, front end managers, and
    telescope operators to define formats and transfer mechanisms.

  – Triaging bug reports and assigning priorities.

  – Creating a unified coding standard.

• Implemented a TAP (Table Access Protocol) service to enable more of
  a direct database access to large astronomy databases. This included
  translating the queries from ADQL to the Postgres, Oracle, Informix
  and SQLite backends and implementing a queue system based on Slurm.
  https://irsa.ipac.caltech.edu/docs/program_interface/TAP.html

• Created a web backend to query a specialized multi-terabyte data
  store containing all of the processed data from the Planck
  satellite. This included a tool to dynamically generate images from
  single pixel observations using various statistical algorithms.
  https://irsa.ipac.caltech.edu/applications/planck/

2005-2012: Lead Software Engineer, Computational Infrastructure for Geodynamics, Caltech

• Created Gamra, a parallel, finite difference, adaptive mesh
  refinement C++ code for modeling deformation during an earthquake.

• In collaboration with a group in Melbourne, Australia, enhanced the
  numerical techniques and physics implemented in Gale, a parallel
  finite element and particle-in-cell code that models mountain
  building, rifting, and subduction in the Earth’s crust.