Myself

Srikrishnaa Jayaraman

Chennai, India

srikrishnaa97@gmail.com

http://www.srikrishnaa.in

English, Tamil, Hindi

Areas of Interest: Data Science, Business Intelligence, Machine Learning, Statistics, Data Visualization, Scientific Computing

Hobbies: Gym, Jogging, TV shows, Movies and Books. Avid cricket fan and a fan of Borussia Dortmund football club.


Certifications

IBM Data Science Professional Certificate [link]


Relevant Courses

Computational Physics (in Python), Probability and Statistics, Statistical Mechanics, Evolutionary Dynamics


Extra-curricular

Organiser of Cryptek '17 - an international online cryptography puzzle solving quiz event.

Secretary of AARSHI (2017) - Dramatics Club of IISER Kolkata. Performed in the play "Wedding Album" by Girish Karnad (2016). Directed the play "Are You Watching Me?" by Tony Frier (2018).

Center topper (Kolkata) of Mimamsa '17, the annual science quiz competition held by IISER, Pune.



Projects

NIFTY-50 Stock Price Prediction [website]

This is an end to end machine learning project to predict the future stock price (30-days) of Nifty-50 stocks from the data set available on kaggle (up to 30-Apr-2021). To predict the stock prices for 30 days from 30-Apr-2021, the stock price for the previous 1000 days was fed into a deep learning neural net consisting of LSTM layers, as sets of 100 days (set 1: day 1 to day 100 stock price to predict the stock price for day 101, set 2: day 2 to day 101 stock price to predict the stock price for day 102, etc). The same neural net architecture was used for all stocks which meant that for some stocks the model performed very well whereas for some stocks, the model performed poorly.

A website was deployed using Flask/MS Azure to display the results obtained from the models. The Google charts library was used to display the charts in an interactive way.


SpaceX Falcon 9 First Stage Landing Prediction [ppt]

SpaceX advertises Falcon 9 rocket launches on its website, with a cost of 62 million dollars; other providers cost upward of 165 million dollars each, much of the savings is because SpaceX can reuse the first stage. Therefore if you can accurately predict the likelihood of the first stage rocket landing successfully, you can determine the cost of a launch.

Data was sourced from SpaceX's REST API and webscraping the wiki article on SpaceX Falcon 9 launches. The data was then filtered, wrangled, stored (in a SQL database), visualized, and analyzed using python libraries such as pandas, seaborn, matplotlib and folium. The exploratory findings from the data was also displayed as a web dashboard using plotly dash.

Different machine learning classification models were used to predict the landing success of Falcon 9's first stage - logistic regressor, support vector classifier, decision tree classifier, and k nearest neighbours classifier. Using GridSearchCV, the best parameters for each model was found and the accuracy score for each model was found using a test set data. The best classifier was logistic regressor with 83.3% accuracy. A presentation was made to share the findings.

This project was done as part of the IBM Data Science Professional Certificate.


Flight Fare Prediction [website]

This is an to end implementation of flight fare price prediction machine learning project. Using the EaseMyTrip website data set on kaggle, I trained a Random Forest Regressor to predict the flight fare prices (in Indian Rupees) between source and destination cities based on a rich array of feature variables such as date and time of arrival and departure, date of search, layovers and the corresponding flight carriers. A fair bit of data preparation was needed due to the variety in the data types of the feature variables. The model was deployed using python Flask module on the MS Azure platform.


Iris Classifier [website]

This was my first end to end project. Using the Iris data set provided in the python sklearn.datasets package, I trained a machine learning model called Support Vector Classifier (sklearn.svm.SVC), to classify the different Iris species based solely on the sepal and petal features of the individual. The model was tested with known labelled data and performed extremely well. The model was finally deployed using python Flask module on the Heroku platform. The whole process was explained step-by-step and recorded on a Jupyter notebook file. I used this mini project to teach and train myself to create more end to end projects.

Technical Strengths

Python

5+ years experience in scientific computation using Python libraries:

Scientific Computation: numpy, scipy

Data Analysis: pandas

Data Visualization: seaborn, matplotlib, folium

Machine Learning: scikit-learn, Tensorflow 2.0

Dashboarding: flask, plotly dash

Webscraping: beautiful soup

Strong knowledge of topics in machine learning and artificial intelligence.

Strong knowledge of machine learning concepts using Tensorflow 2.0 and scikit-learn packages. Have implemented concepts for standalone end-to-end projects.


Web Dev

7+ years experience coding in HTML and CSS. Part of the web dev team for IISER Kolkata college fest 2 years in a row (2016 and 2017). I was in-charge of the website backend - implemented successfully using PHP and MySQL.

Successfully conducted a week-long international online cryptography quiz event - Cryptek (2017). I was part of the quiz team as well as the sole member of the web dev team - designed both front end and back end of the website. Also made a separate quiz website for the pre-event promotional intra-college event SHA256.


MATLAB

Knowledge and experience implementing scientific computations in MATLAB.


Other Programming Languages

Strong knowledge in SQL and querying data.

Knowledge in Java, C++, C, and OOPS concepts and implementations.


Other Software

LaTeX: 3+ years compiling and editing my report summaries on LaTeX, my preferred typesetting software.

MS Office: working knowledge of MS Word, MS Excel and MS Powerpoint.

Operating Systems: Have worked with Windows, MacOS, and Linux systems. Have the ability to switch between operating systems seamlessly. Comfortable with command line syntax of all three operating systems.


Work Experience

PhD - Physics of Living Matter
Feb 2021 - Sep 2021

Principal Investigator: Dr. Anupam Sengupta, Physics of Living Matter Group, University of Luxembourg.

Microbial interaction with its microenvironment determines its activity and fitness and is crucial to ensure the survival of its species. The microenvironment, in itself, is a physical system which responds to collective microbial activity. This soup consisting of microbes and their microenvironment is a non-equilibrium system that provides the recipe for microbial adaptation, microenvironment manipulation, and eventually, the perseverance of life on the planet.

With a thoroughly mechanical approach to studying microbe-microenvironment interactions using tools from biological laboratory (cell culture, microscopy and image analysis) as well as agent-based and coarse grained in-silico models, the goal is to come up with a universal mathematical framework that could predict microbial behaviour, activity, fitness, based solely on the physical properties of the complex environment (density, topology, flow speed, topography, viscosity, etc.) and the morphology of the microbe (shape, aspect ratio, motile/non-motile, etc.).


MS Thesis - Experimental Condensed Matter Physics
May 2019 - June 2020

Ultrafast Charge Carrier Dynamics of Silver Nanowires [pdf]

Principal Investigator: Dr. Kamaraju Natarajan, Ultrafast THz group, IISER Kolkata.

Ultrafast charge carrier dynamics studies in several materials have given important insights into the relaxation dynamics of carriers and possible bandstructure of the material. These studies have led to the effective utilization of these materials in the real world.

Nanostructures are ubiquitous in material sciences due to a high degree of control in the structure during fabrication and their vast scope of applications. At the nanoscale, the charge carrier dynamics vary significantly from the bulk matter counterpart; hence it becomes important to study the ultrafast carrier dynamics of these structures as well. Several studies on metal and semiconductor nanostructures have shown that carrier dynamics are very sensitive to the nanostructure size and shape. This results in a high degree of tunability in the nanostructures for diverse applications. An attempt to study ultrafast carrier dynamics of silver nanowires (AgNW) using pump probe spectroscopy is explained. Numerical simulations of nanoparticle and nanorod pump probe behaviour based on Mie theory and two temperature model were performed.

The simulations (using MATLAB) show that the value of the negative absorption maxima tends to increase when the pump wavelength is decreased and the decay time of the pump probe absorbance signal tends to decrease with pump fluence. Finally, single beam transmitted light microscopy was performed on MoS2 powder sample.


Projects
August 2019 - November 2019
Independent Study Project

Quantum Weak Measurement and its Implementation in Optical Systems [ppt,pdf]

Principal Investigator: Dr. Nirmalya Ghosh, BioNaP group, IISER Kolkata.

A quantum measurement which does not result in the collapse of the wavefunction into one of the eigenstates is known as weak measurement.

Over the course of a semester, I was able to study and provide insights from several different papers to observe weak measurements in optical systems to the BioNAP group in IISER Kolkata. I used python as a teaching/graphical tool in my presentations, to illustrate essence of weak measurement.


May 2018 - July 2018
Max-Planck-Institut Für Kernphysik, Heidelberg

TNSA and Mathematical Modelling of Plasma [pdf]

Principal Investigator: Dr. Adriana Pálffy, Theoretical Quantum Dynamics and Quantum Electrodynamics Group, MPIK Heidelberg

The project involved solving a system of simultaneous partial differential equations in an attempt to monitor the fusion reaction taking place in stars. This experiment was simulated by Dr. Adriana Palffy and Dr. Yuanbin Wu using an existing fluid model for plasma. I succeeded in replicating their results using python.

Teaching Assistantships

Spring 2020

Thermal Physics: Basic thermal physics course for students of physics in their second year

Course Instructor: Dr. Arindam Kundagrami, IISER Kolkata


Autumn 2019

Optics, Electronics, Modern Physics Laboratory: Basic optics and modern physics lab course for students of physics in their second year

Course Instructors: Dr. Nirmalya Ghosh and Dr. Chiranjib Mitra, IISER Kolkata


Education

Indian Institute of Science Education and Research, Kolkata
August 2015 - July 2020

Bachelor - Master Dual Degree in Physics (minor in Mathematics)

CGPA - 8.25 (on a scale of 10)


Chettinad Hari Shree Vidyalayam, Chennai
2015

Class 12 - 96.4%, Indian School Certificate (ISC)


Chettinad Hari Shree Vidyalayam, Chennai
2013

Class 10 - 92.3%, Indian Certificate of Secondary Education (ICSE)