Akhil Ravoori

å adl12w@r.postjobfree.com D Akhil Ravoori µ +91-911*******

Education

Birla Institute of Technology and Sciences Pilani Hyderabad B.E. Electronics and Instrumentation Aug 2017 – Jul 2021

• Overall CGPA: 6.73/10

• CGPA of Last 65 Credits: 7.38/10

• CGPA of All Electives taken in Last 3 Semesters: 7.73/10

• All Elective Coursework in the Last 3 Semesters: Object Oriented Programming, Data Structures and Algorithms, Econometric Methods, Data Mining, Human Computer Interaction(HCI), Marketing Research, Operating Systems, Discrete Mathematics, Selected Topics of Computer Science, Internet of Things, Game Theory, Cinematic Adaptation Relevant Internship Experience

Research Intern Oct 2020 – Now

Centre for Analytical Finance, Indian School of Business Remote, Part-time

• Was one of 10 people selected for this internship out of a pool of 600 applicants

• Worked on 3 projects over the course of this internship:

* Ayushman Bharat and its Impact on the Economic Welfare of India

* West Bengal Legislative Assembly Elections Prediction

* Intrinsic Value of Corporate Social Responsibility (CSR) in India

• Currently using NLP based techniques to analyse company annual reports to formulate a trading strategy for the Indian stock market

Research Intern Jan 2021 – Now

Idea to Startup Labs (ISLabs), Indian School of Business Remote, Full-time

• ISLabs is a research initiative conducted by both the Indian School of Business and Bocconi University to study the decision making of entrepreneurs

• Working on the data generating phase of this large scale research project

• Interviewed over a 100 entrepreneurs of early phase start-ups who are affiliated with ISLabs, to monitor the progress of their businesses

Research Work and Coding Projects

Ayushman Bharat and its Impact on the Economic Welfare of India Python Feb – March 2021

• Structured highly sensitive raw data collected from a Tableau server provided by CRIF High Mark, a credit bureau in India

• Used selenium to automate data collection from official government data sources on the National Family Health Survey conducted in India and COVID-19

• Successfully tested the indifference between control and treatment groups prior to the implementation of the scheme, by performing multiple t-tests on various datasets for rainfall, agricultural production, nightlights, and credit

• Executed Difference-Indifference Regressions between default rates on different types of loans and an interaction variable which defined whether or not the scheme was implemented at a particular point in time for each state in India

• Designed placebo tests to check whether or not the resultant coefficients of the initial regressions were outliers.

• The initially observed coefficients were found to be outliers in the 99th percentile hence reinforcing our findings

• Libraries used: pandas, numpy, linearmodels, selenium West Bengal Legislative Assembly Elections Prediction Python Jan – Feb 2021

• Used fundamental statistical models such as OLS and Logit to run regressions on the data to see if further complex models should be used.

• Used a XGBoost Classifier to estimate probabilities of winning for the ruling party of the nation

• Collected data from official government sources on various factors such as Employment Rates, Crime Rates, Insurance Coverage, etc. to be taken as input into the model

• Libraries used: pandas, numpy, sklearn, xgboost, matplotlib Intrinsic Value of Corporate Social Responsibility (CSR) in India Python Oct – Dec 2020

• Researched the effect of the CSR Mandate of 2013 on the intrinsic motivation of companies to expend towards CSR related activities

• Used selenium to collect data on CSR expenditures of approximately 12,000 companies and collected tweets of 5000 companies using a Python based library called Twint and Tweepy

• Pre-processed the data to be ready for fitting regression models and textual analysis

• Conducted topical analyses on the tweets and observe how companies reduced advertising their CSR related activities and increased advertising of their products

• Built multiple dictionaries to filter tweets into two main categories such as: CSR and Product.

• Gained insight on the above by analysing the frequencies of such tweets being posted prior to the mandate and after.

• Used cosine similarity as a metric to analyse the textual similarities between sections on CSR of company annual reports

• Checked if these sections within the treatment group’s annual reports are significantly different after the mandate was imposed, as compared to prior using this metric.

• Libraries used: nltk, pandas, tweepy, twint, selenium Deeper Dive into Social and Print Media - A NLP based analysis Python Jan – Dec 2020

• Collected a dataset of 45k tweets as well as 7k online traditional media articles over a two week period after the Abrogation of Article 370

• Used IBM Watson’s Tone Analyser API to perform unsupervised sentiment analysis on the text thereby automating the labelling of the dataset for our models to train on

• Analysed and created data visualisations for the trends of the sentiment scores obtained from the Tone Analyser for both the social media texts and the traditional media articles

• Trained Neural Network models (CNN’s and LSTM’s) to perform sentiment analysis on the textual data that was collected

• Libraries used: pandas, numpy, IBM Watson’s API, keras, matplotlib Extracurriculars

• Was a part of the College Basketball Team

• Cleared the ABRSM Level 4 Piano Exam with Merit

Skills

Languages: Python, C/C++, Java, JavaScript, R

Developer Tools: Jupyter Notebooks, VS Code, Spyder, Android Studio Human Languages: English, French, Telugu, Hindi

Contact this candidate