Algorithmic Trading

Devised trading strategies leveraging sockets and threads to enhance market responsiveness and execution efficiency

Developed a dynamic market platform that matches traders based on optimal prices, enhancing overall market performance and has robust measures to identify and prevent arbitrage opportunities

Implemented median trading and statistical arbitrage, to exploit market anamolies and drive consistent returns

Option Pricing Models

Immersed in the dynamic realm of Stock Markets and Options Trading, comprehending various Option Strategies

Implemented the Black-Scholes model, the Binomial model, and Monte Carlo simulations using Python libraries

Evaluated the precision and performance of the Black-Scholes Model by applying it to real-world data sourced from the National Stock Exchange (NSE) markets and achieved an accuracy of 90 % by optimizations

Economics Meets Machine Learning

Formulated and implemented various economic problems as Markov Decision Processes in the Gym framework

Employed a combination of Bandit algorithms and Reinforcement Learning algorithms, known for their adaptability and learning capabilities, to address complex matching markets, auction dynamics, and allocation problems

Modeled stock exchange as a double auction, which incorporated market sentiment and the individual objectives of users and conducted simulations involving a diverse group of over 100 participants to assess the model’s efficacy

Computer Architecture Implementation

Utilized the champsim simulator to implement and analyze stream and IP stride prefetchers, while evaluating the effectiveness of LRU, FIFO, LFU and BIP replacement policies based on IPC and accuracy metrics

Designed a VHDL circuit which encodes musical chords, achieving conversion of 8-bit binary notes into chords

Implemented Heap Sort, Merge Sort and Binary Search algorithms in the MIPS and x86 Assembly Language

Reinforcement Learning

Completed an extensive long reading project on Reinforcement Learning (main reference: Sutton & Barto)

Investigated Dynamic Programming, Monte Carlo Methods, n-step bootstrapping, Temporal Difference learning and on-policy methods with their applications and implemented all the algorithms in Python

Image Segmentation

Implemented Gaussian Mixture Model (GMM) with Saliency Map for accurate image segmentation

Achieved better accuracy and reduced computational cost over Markov Random Field and mean template based GMM

Mathematics of Derivative Pricing

Achieved proficiency in mathematical models used in derivative pricing through comprehensive coursework

Acquired knowledge about Derivatives such as Futures and Options, with a focus on trading strategies

Gained in-depth understanding of Black-Scholes model, Binomial model, and Greeks for risk management

Python Web Crawler

Designed a sophisticated Web Crawler equipped with the ability to recursively extract all the hyperlinks of a webpage and generating a comprehensive graph showcasing various link types, with user-defined recursion levels

Harnessed the power of Python libraries to visualize this wealth of data, resulting in an elegant representation

Produced an extensive report and incorporated a user-friendly Makefile for compilation of the code and report

Line-Follower Robot

Designed a robot to autonomously track lines on inclined surfaces and successfully unload cargo at the destination

Fostered teamwork and management skills by navigating through intricate electrical and mechanical concepts

Implemented algorithms for real-time line detection and precise control, significantly augmenting the robot’s ability to navigate complex terrains with exceptional accuracy thereby greatly improving its operational effectiveness

Alien Invasion Game

Developed a customized version of Alien Invasion Game aiming for an engaging gameplay experience

Added layers of excitement and depth to the gaming experience by allowing the players to shoot advancing extraterrestrial invaders, encountering three unique types, each with individually tailored point rewards

Enhanced gameplay by introducing escalating difficulty levels, with each new level increasing game speed and point rewards for aliens and implemented a high-score tracking system that saved and retrieved player high scores