Internships, Projects and Collaborations

I have worked/am working with Scientists from Institutes which include: California Institute of Technology (Caltech), Raman Research Institute (RRI, India), NCRA-TIFR (India), University of Delaware, University of Moncton, Indian Institute of Technology, Indore (India), University of Calcutta (India), Centre for Development of Advanced Computing (CDAC, India). I have participated in diverse projects and collaborations, with on-site/remote experience at several leading facilities worldwide like:

Array of Precision Spectrometers for the Epoch of Recombination (APSERa)
Calcutta University - Stratosphere Troposphere Radar (CU-STR)
Echelle SpectroPolarimetric Device for the Observation of Stars (ESPaDOnS)
Giant Metrewave Radio Telescope (GMRT).

My work is jisted below for a short read, but they are linked to my Major Projects page where you can find detailed information on the project background, work done, or publications produced. Please drop me a mail at arjunghosh@ieee.org if you wish to know more about any of my work.

Ongoing Projects

APSERa Experimentation

On-site, Funded by RRI (May 2024 - ongoing)

Guide: Dr. Mayuri S. Rao and Keerthipriya Sathish, Raman Research Institute

Instrumentation work with APSERa (Antenna Array to probe Recombination ripples from Epoch of Recombination)

Developed software for APSERa, including Web app for system monitoring and control.
Optimized system performance through several runs, achieving temperature stability at 75K, 77K, 120K and conducting pressure variation studies for leak studies.
Modeled the cryocooler system and estimated its parameters from its time response curves.
Designed and integrated a thermal shroud into APSERa to improve heat dissipation.
Conducted antenna and radome simulations to optimize performance and minimize edge diffraction.
Prepared an experimental deployment plan for APSERa.
Developed and validated a simulation pipeline to analyze RF chain non-linearities and systematics, using MATLAB and CST. Journal in preparation

MACDA (MCMC Assisted Coupler Designer Analyzer): A new design method for Hybrid Couplers.

Remote, Un-paid voluntary Internship (Dec 2023 - Present)

Guide: Dr. Ritoban Basu Thakur (NASA-Jet Propulsions Lab and Caltech)

Quadrature Hybrid Couplers (QHDC) are critical components in RF, mm-wave, and sub-mm wave astronomical instrumentation, where wideband performance with minimal passband ripple is essential. Traditional designs have been limited to 5-sections at most, due to computational limitations. In this work, we introduce a new analytical technique to design couplers with larger sections and improved performance. We do this by employing a Markov Chain Monte Carlo (MCMC) based solver. By defining a likelihood function based on S-parameter equations and incorporating physical priors, we derive optimized impedance values that enhance bandwidth beyond what is reported in the literature. Our flexible pipeline allows efficient tuning of the coupler design. The results demonstrate fractional bandwidths that reach 1.0 for a 9-section coupler, substantially outperforming previous designs. Statistical analysis and convergence tests confirm the robustness of our approach. Read our preprint (Click). Journal in preparation

Release Plan for MACDA: Currently we are working on upgrades to MACDA to incorporate arbitrary coupler design subroutines as well, following which a release plan will be designed in a follow-up paper.

Past Projects

Active Mode experimentation with CU-STR

On-site, Part of coursework (Sep 2023 - May 2024)

Guide: Dr. Ashik Paul (Head, IRPE, CU) · Dr. Souvik Majumdar

Final year BTech theses: Studying Atmospheric Gravity Wave in Eastern region of India with CU-ST Radar for the first time. Read about this here.
Worked and operated the RADAR for wind profiling of the atmosphere in height range of 1.5 Km to 23 Km.
Performed Moon Echo Experiment.

CU-ST RADAR in Astronomy (Passive)

On-site, Partly funded by IIT-I (June 2022 - Jan 2024)

Institute of Radio Physics and Electronics, University of Calcutta, India

Guide: Dr. Ashik Paul; Dr. Abhirup Datta; Dr. Narendranath Patra

Time Domain observation of Pulsars.
Performed tracking experiments with the CU-ST Radar using Sun, Virgo A, Taurus A, Cassiopeia A as Source.
Worked on characterizing the antenna with Noise, Gain, Peak Power Profiles of for Astronomical Observations.

More details on passive mode operations here

Specpolflow module functionalities

Remote, Unpaid voluntary work (Sept 2023 - Jan 2024)

Guide: Dr. Veronique Petit, Dr. Christi Erba, Dr. Colin Folsom .

Added functionalities and GUI to existing stellar analysis tools: specpolflow and LSDpy by Colin Folsom and Christi Erba

Github Links: Application developed by me, LSDpy, specpolFlow

MITACS Globalink Research Internship

On-site, Funded by MITACS[1] (Canada) - AICTE[2] (India) - SICI[3] (India) (July 2023 - Sept 2023)

University de Moncton, Moncton, New Brunswick, Canada

Guide: Dr. Victor Khalack

Studying Chemically Peculiar stars using the ESPaDOnS

Analyzed available spectra of CP stars and derived their radial velocity, mean longitudinal magnetic field and stellar rotation. Developed pipelines for Least Square Deconvolution and Mean Longitudinal Magnetic Field Analysis.

Derived Mean Longitudinal Magnetic Field (Bz) and Vsin(i) for CP stars by fitting their Balmer line profiles in the spectra provided by ESPaDOnS at the CFHT telescope.
Created plots of Magnetic Field Variability with Phase of Star, for improving value of stellar rotation.

More details here

[1] MITACS (Mathematics of Information Technology and Complex Systems)

[2] AICTE (All India Council for Technical Education)

[3] SICI (Shastri Indo-Canadian Institute)

ML Applications for COVID-19

(June 2022 - December 2022)

Centre for Development of Advanced Computing, Pune

Guide: Dr. Anil Kumar Gupta

Visualization for input processed data and output result data for an automated machine learning model.

Created a linear regression based model to detect COVID-19 from the patient's cough sounds, studying the correlation between various spectral features of cough sound and the viral infection. Our model gave us an accuracy of 85%.

Small projects

COVID Detection from Lung Cough Noise using Machine Learning May 2022 - September 2022 (B.Tech / Team Project) University of Calcutta:

We have successfully made a working model to diagnose COVID-19, efficiency around 90%.

Link to project

2. Antenna Design using HFSS April 2022 (B.Tech / Faculty: Prof. Santanu Mondal) University of Calcutta

Designed a Microstrip Dipole Patch Antenna at 2.4 GHz using CST Microwave Studio for applications in Small Scale communication links.