WorldQuant | Quantitative Research Consultant | Remote

July 2024 – Current

• Developed over 15 trading strategies (Alphas) using quantitative finance models for U.S. and China markets.
• Conducted 10-year market data backtesting, analysing performance metrics such as Sharpe ratio, turnover, and returns.
• Selected from the top 1% of WorldQuant Challenge participants for exceptional performance in quantitative research.

Bloomberg L.P. | Software Engineer, Intern | Dublin, Ireland

May 2023 – Sep 2023

• Added health endpoints with advanced features for improved monitoring and incorporated time control features to boost efficiency. Rectified critical service bugs related to null errors, job retries, and time discrepancies, ensuring data integrity and system reliability.
• Automated file generation with FreeMarker templates, streamlining the transition from over 10 ‘.dp’ to JSON and dppfest files.
• Employed rigorous unit testing on over 15+ services, employed Antlr for file parsing, and resolved CSV duplicate entries to ensure data accuracy and code reliability. Leveraged Spring internal functions, including retry templates, for bug fixing increasing efficiency by 5%.
• Enhanced Kubernetes service connectivity by utilised JavaScript and optimised the UI achieving 85% seamless integration, managed deployment pipelines with Jenkins for over 50 deployments.
• Leveraged Knowledge: Java, Python, Spring Boot, JavaScript, FreeMarker, Antlr, Junit, Jenkins, Bitbucket, Jira, Swagger, Ansible, Gradle, Apache Maven, Confluence.

Vantage Circle | Software Engineer, Intern | India

Jul 2021 – Sep 2021

• Worked on hosting an internal expense management tool on GitHub and improving the integration between the backend and database for the tool.
• Ran internal queries on Postgres Database with Python’s API for Postgres. Researched about Apache Calcite and where it would be suitable to be used for internal company requirements.
• Leveraged Knowledge: Git, GitHub, SQL, Calcite, Postgres.

Software Engineer, Intern | SYSWARE Infotech Private Limited

June 2020 – August 2020

• Worked on developing ML models to analyze the current and archived incident tickets in the internal incident tool.
• The models help categorize and find patterns on the incidents, which further aided the incident team to improve their incident clearance and management.
• Leveraged Knowledge: Python, ML Libraries, and packages.

Thesis project: Impact of News Sentiments on Stock Prediction Model during periods of Global Crisis using Sentimental Analysis

• Python
• Long Short-term Memory (LSTM)
• Prophet · BERT (Language Model)
• Keras
• Transformers
• Machine Learning
• Data Visualization
• Data Analysis
• Data Structures
• Deep Learning

For my thesis project, I analyzed the impact of global crises on stock markets, focusing on leading indices such as the Dow Jones and S&P 500 during events like COVID-19 and geopolitical disruptions. By integrating sentiment analysis from financial news headlines using FinBERT with LSTM and Prophet models, I achieved an average increase in MAPE of 0.89% for LSTM and 0.50% for Prophet when sentiment scores were included. I developed both models, demonstrating LSTM's superior accuracy, particularly in volatile market conditions, with an average MAPE of less than 5%. The project offered valuable insights into market resilience, and I visualized the results through a real-time Streamlit dashboard for interactive stock market analysis.

Biomedical Imaging AI for Cancer Tumour Segmentation & Classification

• Python
• Convolutional Neural Networks (CNN)
• Machine Learning
• Data Visualization
• Data Analysis
• Data Structures
• Deep Learning

I developed a Convolutional Neural Network (CNN) model for image classification, achieving an accuracy of 86.04%, effectively categorizing images from a diverse dataset. Additionally, I implemented a machine learning model for breast cancer tumor segmentation and classification using ultrasound scans. This model attained an accuracy of 83.69% for classification and an F1 score of 0.922 with 96.70% accuracy for segmentation, demonstrating its strong performance in medical imaging tasks.

Investment Strategy Analysis Dashboard

• Python
• Streamlit
• Pandas
• NumPy
• Altair

I developed an advanced investment analysis dashboard using Streamlit, enabling real-time financial data manipulation and visualization. The dashboard incorporates robust Python scripts to calculate key financial metrics like Geometric Mean Return, Volatility, VaR, and CVaR, providing comprehensive risk assessment tools for investment decision-making. I employed complex financial models to compare portfolio performance, while also integrating Excel data handling for seamless management. Using Altair, I created intuitive visualizations of investment metrics, and enhanced user engagement with interactive widgets, allowing for customization of parameters. Backend calculations were optimized with Pandas and NumPy to ensure high-performance real-time data processing.

Interactive Map Application

• JavaScript
• Leaflet.js
• Express.js
• Node.js
• Multer
• Cascading Style Sheets (CSS)
• HTML5

I developed a dynamic web-based interactive mapping application using Leaflet for front-end map features and Express.js for the back-end server setup. The application allows users to interact with the map by adding or removing markers and creating routes with polylines, while also visualizing geographic data. I integrated file upload capabilities using Multer, enabling users to attach media files like images or videos to specific locations on the map. Advanced geospatial editing features were implemented using Leaflet’s Drawing Tools, and I optimized route calculations through Leaflet Routing Machine. Additionally, I enhanced the user interface with custom CSS for responsive design and ensured server reliability with efficient routing and error handling using Node.js and Express.

Stock Predictor & News sentimental Analysis

• Python
• yfinance
• pandas_datareader
• TextBlob
• NLTK
• Flair
• transformers
• matplotlib
• plotly
• Keras
• LSTM
• Streamlit

Developed a financial analysis tool in Python, integrating real-time stock prices (via yfinance and pandas_datareader) and news articles for insights using NewsApi. Conducted sentiment analysis on stock-related news using TextBlob, NLTK, Flair, and transformers; visualised correlations on matplotlib and plotly. Built a Keras-based deep learning model / machine learning with LSTM for stock price prediction, achieving less than 5% error after 100+ iterations. Deployed a Streamlit web dashboard for real-time viewing, enhanced with error handling and integrity checks. Increased data cohesiveness by aligning dates between stock prices and news data, prioritising accuracy, and robustness.

Decentralised Chat App

• Python
• UDP
• SHA-256
• OpenVPN

Developed a decentralised P2P network on Pi platforms for resilient communication. Utilised UDP for lightweight, low-latency transmission and integrated SHA-256 for secure authentication. Implemented NAT hole-punching with a rendezvous server, allowing seamless peer communication. Gained proficiency in Python socket programming and applied OpenVPN for secure server access.

Dublin Bus Route Finder

• C
• Dijkstra's algorithm

Dublin-Bus-Root-Finder is a C-based application that identifies the shortest bus route between two specified stops in the Dublin Bus network. Using graph structures and Dijkstra's algorithm, it represents the network where stops are vertices and routes are edges. The application can load data from CSV files and offers a clear output of the shortest route.