Projects Portfolio 🚀
Welcome to my Projects Portfolio! Here, you'll find a comprehensive overview of my technical projects, divided into two main categories for easier navigation: Quant Projects and Machine Learning & Data Engineering Projects
1. Quant Projects 📊
1.1 Nasdaq Closing Price Prediction Challenge
Description: Developed a model to predict closing price movements for Nasdaq-listed stocks during the final ten minutes of trading, leveraging order book and auction data.
Project Link: Click Here
Challenges and Considerations:
- Combining Order Book and Auction Data: Participants should effectively merge information from both the order book and closing auction data to make accurate predictions.
- Supply and Demand Dynamics: Understanding and modeling supply and demand dynamics is crucial for predicting closing price movements.
- Trading Opportunities: The models should identify potential trading opportunities in the final ten minutes of the trading session.
Inspiration: Based on Optiver’s problem statement.
1.2 Realized Volatility Prediction
Description: Created models to predict short-term stock market volatility across various sectors, critical for financial product pricing.
Project Link: Click Here
Challenges and Considerations:
- Short-Term Volatility Forecasting: Accurately predicting short-term fluctuations in stock prices over 10-minute periods is a significant challenge.
- Pricing Implications: The models developed in this competition have important implications for the pricing of options and other financial products.
Inspiration: Based on a problem statement by Two Sigma.
1.3 Jane Street Market Prediction Challenge
Description: Developed a multi-model ensemble combining LightGBM and a custom cost awareness function to enhance market prediction accuracy, resulting in an 8% increase in the sensitivity score.
Project Link: Click Here
Challenges and Considerations:
- Multi-Model Ensemble: Combining different models to create an effective ensemble that improves predictive accuracy is a key challenge.
- Custom Cost Awareness Function: Designing a cost awareness function that can effectively optimize trading decisions based on market conditions.
- Real-Time Predictions: Utilizing a time-series API for real-time predictions to handle high-frequency trading data and make timely trading decisions.
Inspiration: Based on Jane Street’s problem statement.
2. Machine Learning and Data Engineering Projects 🤖
2.1 Airlines Management Smart System
Project Link: GitHub
About: Enhanced the airline industry's DBMS for storing, organizing, and retrieving comprehensive data segments.
2.2 Stock Predicting Model using RNN during Bull Market
Project Link: GitHub
Summary: Demonstrated building an RNN model with LSTM cells to predict S&P500 index prices, employing TensorBoard for model tracking.
2.3 Credit Risk Model on Machine Learning and Prediction
Project Link: GitHub
Summary: Developed a model for credit risk assessment based on applications from UAE's commercial banks.
2.4 ARS_Cheetah.V2
Project Link: GitHub
Summary: Trained the Half-Cheetah model using the Augmented Random Search algorithm in the MUJOCO environment.
1.1 Nasdaq Closing Price Prediction Challenge
Description: Developed a model to predict closing price movements for Nasdaq-listed stocks during the final ten minutes of trading, leveraging order book and auction data.
Project Link: Click Here
Challenges and Considerations:
- Combining Order Book and Auction Data: Participants should effectively merge information from both the order book and closing auction data to make accurate predictions.
- Supply and Demand Dynamics: Understanding and modeling supply and demand dynamics is crucial for predicting closing price movements.
- Trading Opportunities: The models should identify potential trading opportunities in the final ten minutes of the trading session.
Inspiration: Based on Optiver’s problem statement.
1.2 Realized Volatility Prediction
Description: Created models to predict short-term stock market volatility across various sectors, critical for financial product pricing.
Project Link: Click Here
Challenges and Considerations:
- Short-Term Volatility Forecasting: Accurately predicting short-term fluctuations in stock prices over 10-minute periods is a significant challenge.
- Pricing Implications: The models developed in this competition have important implications for the pricing of options and other financial products.
Inspiration: Based on a problem statement by Two Sigma.
1.3 Jane Street Market Prediction Challenge
Description: Developed a multi-model ensemble combining LightGBM and a custom cost awareness function to enhance market prediction accuracy, resulting in an 8% increase in the sensitivity score.
Project Link: Click Here
Challenges and Considerations:
- Multi-Model Ensemble: Combining different models to create an effective ensemble that improves predictive accuracy is a key challenge.
- Custom Cost Awareness Function: Designing a cost awareness function that can effectively optimize trading decisions based on market conditions.
- Real-Time Predictions: Utilizing a time-series API for real-time predictions to handle high-frequency trading data and make timely trading decisions.
Inspiration: Based on Jane Street’s problem statement.
2.1 Airlines Management Smart System
Project Link: GitHub
About: Enhanced the airline industry's DBMS for storing, organizing, and retrieving comprehensive data segments.
2.2 Stock Predicting Model using RNN during Bull Market
Project Link: GitHub
Summary: Demonstrated building an RNN model with LSTM cells to predict S&P500 index prices, employing TensorBoard for model tracking.
2.3 Credit Risk Model on Machine Learning and Prediction
Project Link: GitHub
Summary: Developed a model for credit risk assessment based on applications from UAE's commercial banks.
2.4 ARS_Cheetah.V2
Project Link: GitHub
Summary: Trained the Half-Cheetah model using the Augmented Random Search algorithm in the MUJOCO environment.
For inquiries or collaborations, feel free to reach out to me at adityasaxena@g.harvard.edu 📧