Boaz Blankrot
Experience
Senior Software Engineer, Machine Learning
2022–Present
Machine Learning Engineer
Virta Health
2021–2022
Artificial Intelligence Fellow
Insight
2020
Addressed the business problem of monitoring transcribed audio ads for a leading audio intelligence platform with a brand and industry classifier using named-entity recognition
Utilized natural language processing techniques and trained a convolutional neural network (CNN) on AWS to achieve 85% brand recognition, and 20% recognition of previously-unseen brands
Research Assistant
Vienna University of Technology
2016–2020
Developed an efficient method for analysis and optimization of dielectric nanostructures consisting of hundreds to thousands of particles, such as photonic crystals and metamaterials, and implemented it as an open-source Julia package.
Vice Chairman and Web Developer (volunteer)
re:member, Vienna, Austria
2016–Present
Vice chairman on board of directors for a nonprofit organization, leading groups of volunteers in both physical activities and digital data entry events. Constructed website with tailored data entry forms using PHP/MySQL/JS.
Researcher and Teaching Assistant
Technion—Israel Institute of Technology
2012–2015
Applied the Fast Multipole Method to a computational electromagnetics method (the Source-Model Technique), thereby reducing solver complexity from O(N3) to O(N2).
Taught two advanced undergraduate courses, Design of Active Filters and Microwaves, for a total of 4 semesters and over 200 students. Designed and graded weekly work assignments and final exams.
Student System Engineer
Zoran, Haifa, Israel
2010–2011
Collaborated on a LabVIEW-based test automation suite for set-top boxes, integrating HDMI, power, analog video, and S/PDIF tests in hot and cold environments.
Education
Ph.D. in Applied Mathematics, Vienna University of Technology
2016–2020
Dissertation: Deterministic and robust optimization of complex nanoscale photonic structures
M.Sc. in Electrical Engineering, Technion—Israel Institute of Technology
2012–2016
Thesis: A Fast Source-Model Technique for Many-Scatterer Problems in Electromagnetics
B.Sc. in Electrical Engineering (cum laude), Technion—Israel Institute of Technology
2008–2012
Specializing in: Waves and Optical Communication, Signal Processing, Microelectronics
Projects
This project uses a TensorFlow CNN to predict the rotation angle of arbitrarily rotated images from the MNIST dataset. The model was trained in a Google Colab notebook, after which I deployed the model to Heroku using FastAPI. In the app, users can choose random images from the test set, apply arbitrary rotation, and see how accurate the trained model is at predicting the rotation angle.
A Julia package for solving large-scale electromagnetic scattering problems in two dimensions; specifically, those containing a large number of penetrable smooth particles. Provides the ability to optimize over the particle radius and rotation efficiently for various multi-wavelength design problems.
Eyelid Motion Monitor
Technion—Israel Institute of Technology
2012
Several neurological diseases and eyelid disorders can be diagnosed by eyelid movement patterns. The goal of this project was to develop a device for monitoring eyelid movement. Our contribution was writing the low-level C code for the device, including data acquisition, as well as a LabVIEW GUI that enables physicians to analyze the recorded data.
Winner of the Oz Moses Foundation Award for Best Project.
Optical Comb Generation Using Mach-Zehnder Modulator
Technion—Israel Institute of Technology
2012
The purpose of our project was to create a laser comb with as many pulses as possible, while maintaining a relatively flat spectrum. This was accomplished in a laboratory using a phase modulator and Mach-Zehnder Modulator in tandem.
Winner of the Seiden Family Prize for Multidisciplinary Undergraduate Student Projects in Optoelectronics, Microelectronics and Nanosciences.
Research
Publications:- B. Blankrot and C. Heitzinger, “On the robust optimization of photonic structures for asymmetric light transmission,” Submitted for publication, Jul. 2020.
- B. Blankrot and C. Heitzinger, “Efficient computational design and optimization of dielectric metamaterial structures,” IEEE Journal on Multiscale and Multiphysics Computational Techniques, vol. 4, pp. 234–244, Oct. 2019.
- B. Blankrot and C. Heitzinger, “Design of aperiodic demultiplexers and optical diodes by optimizing photonic crystals,” OSA Continuum, vol. 2, no. 7, pp. 2244–2252, Jul. 2019.
- B. Blankrot and C. Heitzinger, “ParticleScattering: Solving and optimizing multiple-scattering problems in Julia,” Journal of Open Source Software, vol. 3, no. 25, p. 691, May 2018.
- B. Blankrot and Y. Leviatan, “FMM-Accelerated Source-Model Technique for Many-Scatterer Problems,” IEEE Transactions on Antennas and Propagation, vol. 65, no. 8, pp. 4379–4384, Aug. 2017.
- B. Blankrot and C. Heitzinger, “Efficiently optimizing inclusion rotation angle for maximal power flow,” presented at 14th International Conference on Mathematical and Numerical Aspects of Wave Propagation (WAVES), 2019, Vienna, Austria.
- B. Blankrot and C. Heitzinger, “Automated Design of Photonic Crystal Demultiplexers,” in Proc. of the 12th International Congress on Artificial Materials for Novel Wave Phenomena (Metamaterials), 2018, Espoo, Finland.
- Minisymposium Organizer, “Numerical Methods for Photonics, Optics, and Metamaterials,” at SIAM Annual Meeting, 2018, Portland, OR.
- B. Blankrot and C. Heitzinger, “Integral Equation Solver for Metamaterial Design,” presented at SIAM Conference on Computational Science and Engineering, 2017, Atlanta, GA.
- B. Blankrot and C. Heitzinger, “Multiple scattering approach for dielectric metamaterial analysis,” presented at Austrian Numerical Analysis Day, 2017, Salzburg, Austria.