Hello there! I'm Raavi

Research

1. Algorithmic Construction of Lyapunov Functions

As a part of my summer undergraduate reseach project, I worked on presenting a novel numerically tractable technique for synthesizing Lyapunov functions for equilibria of nonlinear vector fields. Corresponding to an isolated equilibrium point of a given vector field, a selection is made of a compact neighborhood of the equilibrium and a dictionary of functions in which a Lyapunov function is expected to lie. Then an algorithmic procedure based on [DACC22] is deployed on the preceding neighborhood-dictionary pair and charged with the task of finding a function satisfying a compact family of inequalities that defines the behavior of a Lyapunov function on the selected neighborhood.

2. Hallucinations of LLMs - Bachelor's Thesis

I am presently engaged in collaborative research on the topic of Large Language Model (LLM) hallucinations with Prof. Ganesh Ramakrishnan, in conjunction with Adobe. Our work focuses on comprehending the underlying factors that lead to hallucinations in LLMs, devising methods for their identification, and implementing mechanisms to alert instances where LLMs produce inaccurate outcomes. Our specific goal is to establish a robust framework applicable across various tasks, minimizing the need for extensive constraints, data, and computational resources. This endeavor aims to attain optimal results while using limited resources.

Projects

Machine Learning & Reinforcement Learning

1. Bandit Optimization: Report and Presentation
2. Cocktail Party Problem (Audio Source Separation): Report and Presentation
3. Stein's unbiased risk estimate as MSE: Report and Presentation
4. Factors affecting student alcohol consumption: Report
5. Optimal Policy for last wicket of cricket: Report

Electronics

1. Radio Frequency Identification: Report
2. Autonomous Underwater Vehicle: Technical Design Report

Miscellaneous

1. Viterbi Decoding: Presentation
2. An Introduction to Number Theory and Cryptography: Report

Teaching

I have had the wonderful opportunity to be a Teaching Assistant (TA) for 120 + freshmen and sophomores at IIT Bombay on four distinct occasions. My experience about one of them can be found in this blog curated for the Electrical Engineering Department Academic Program, IITB.

1. MA108: Differential Equations - I
2. MA207: Differential Equations - II
3. MA111: Calculus

Internships

Piramal Capital and Housing Finance Limited

During the summer of '23, I had the wonderful opportunity to intern at Piramal Finance. My main focus was on enhancing the credit risk model by effectively categorizing financial transactions from bank statements. The approach I devised involved a well-structured three-stage pipeline. The first stage employed a rule engine to identify and exclude unconventional categories like Bounced Transactions and Dividends. The second stage, which I termed the "Company Extractor," was responsible for identifying the companies mentioned in the transactions. This step was crucial for understanding the domain associated with each company, which enabled accurate tagging. For instance, if the transaction involved a company like SpiceJet, it was appropriately tagged as "Travel" based on the company's domain. In the final stage of the pipeline, I implemented a sophisticated language model with the ability to comprehend non-colloquial keywords, including certain Hindi terms. This feature was utilized to further refine the transaction tags, enhancing the overall accuracy of the credit risk model.

FinIQ Consulting

Starting out, I interned at FinIQ Consulting as a Quantitative Analyst where I designed an optimized pricer for Target Redemption Forward (TRF) structured products.

CV

Contact

• Linkedin
• Gmail
• GitHub

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';