Internship: Characterization of nanoparticles using machine learning

Internship

Characterization of nanoparticles using machine learning

Nanoparticles are of increasing importance for health care (nanomedicine), semiconductor industry (quantum dots) and the environment (nanoplastics). One way to measure nanoparticles is by using flow cytometry, a method where a laser beam illuminates a particle and the light scattering of the particle is monitored. Particle properties are  generally determined using Mie theory which, given the properties of the particle (e.g., radius and refractive index), determines how the scattering pattern will be when the particle is hit by a laser of given properties (e.g., wavelength).

In our measurement setup the challenge is to solve the inverse problem, i.e., given the scattering pattern determine the radius and refractive index of the particle that generated it. The inverse problem is ill-posed and one way to solve it is by using machine learning (ML). However, the performance of the ML approach needs to be benchmarked against other traditional methods.

As part of this internship the candidate will

• Perform a literature review on available methods (excluding ML) to solve the inverse problem in Mie theory
• Select and implement at least one approach identified in the literature review
• Fine tune and benchmark an in-house ML model designed to solve the inverse problem
• Evaluate the uncertainties associated with the predictions of the approach from the literature and with the predictions of the ML model

The work will be carried out at VSL (Delft) within the Data Science and Modelling group and in collaboration with the Length group. The ideal candidate has a good understanding of inverse problems, numerical analysis, and ML. Programming experience (preferably in Python) is required.

Internship term: Six months
Location: Hybrid with at least 2 days per week at our office in Delft
Allowance: A monthly allowance of € 615,- (gross) is applicable