Servus! I am Sandeep, currently working as Staff Software Engineer at Arriver Software GmbH.

Professional and research history

Before starting current job in 2020, I did a short postdoc at Technical University of Ilmenau in Germany, where I focused on a turbulence related problem and developed novel data-driven methods based on machine learning and reduce-ordered modelling. During this sojourn at Ilmenau, I have worked with deep neural network, recurrent neural network and reservoir computing where I heavily rely on Python, MATLAB, Keras and associated libraries.

I did my PhD at University of Stuttgart where my prime aim was to develop data-driven models for heat transfer prediction in a new-type of power plants. As I was missing the experimental data, so I did several highly accurate and computationally expensive (1 simulation takes approx 400,000 CPU-hours!) simulations on supercomputers using up to 50,000 CPUs at once. Some of the data from these numerical experiments were released for benchmarking of new methods. After having data in hand, I worked on some regression based models and then explored the machine learning based methods including decision tree, random forest and deep neural network. My work was not only limited to model development but I also have to perform data analytics for the high-fidelity data coming out from the simulation. During data analytics part, I developed different toolkits within MATLAB and Python to perform variety of task such as probability analysis, eigenvalue decomposition, correlations, and statistical analysis.

I hold a Master of Technology degree from IIT Delhi in Energy Studies, where I had unique opportunity to dive deep into the energy resources modeling from technical, economical and environmental perspective. I did my Bachelor of Technology in Mechanical Engineer from RTU Kota.

Curious?

If you have any question, suggestion about my work, then please feel free to drop me a mail! I am also open to any collaboration within my domain of expertise.