In collaboration with WorldQuant and Udacity, I've completed a comprehensive program focusing on quantitative finance and AI applications in trading. Throughout this program, I've acquired advanced skills in financial data analysis, model building, and algorithmic trading strategies. Equipped with this expertise, I am ready to contribute effectively to the dynamic environment of quantitative asset management.
Quantitative Analysis Proficiency: Mastered quantitative analysis techniques including data processing, trading signal generation, and portfolio management. Advanced AI Applications: Utilized machine learning algorithms such as natural language processing, recurrent neural networks, and random forests to generate trading signals and evaluate strategies. Real-World Project Experience: Successfully completed 8 projects demonstrating expertise in applying quantitative finance principles to real-world scenarios. Collaboration and Communication: Worked with industry experts and peers, fostering collaborative skills and effective communication in presenting findings and strategies.
Project 1: Trading with Momentum This project involves independently implementing and testing a trading strategy to assess its potential profitability. You'll be provided with a set of stocks and a specified time range. Your task is to generate a trading signal based on a momentum indicator, compute the signal for the given time range, and apply it to the dataset to project returns. Finally, you'll conduct a statistical test on the mean returns to determine if the signal exhibits alpha. The dataset utilized will be Quotemedia's end-of-day data.
Project 2: Breakout Strategy In this project, you'll implement a breakout strategy, which includes identifying and handling outliers. Through statistical analysis, including histogram and P-Value assessments, you'll evaluate the strategy's potential profitability. Furthermore, you'll run scenarios with and without outliers to determine their impact on the trading signal's effectiveness. Similar to Project 1, Quotemedia's end-of-day data will be utilized.
Project 3: Smart Beta and Portfolio Optimization Here, you'll construct a smart beta portfolio and compare its performance against a benchmark index. By calculating tracking error and optimizing weights through quadratic programming, you'll build and rebalance the portfolio, assessing its performance through turnover analysis. The optimal rebalancing frequency will also be determined. The dataset employed will be Quotemedia's end-of-day data.
Project 4: Alpha Research and Factor Modeling This project entails building a statistical risk model using Principal Component Analysis (PCA) and incorporating five alpha factors to construct a portfolio. Evaluation metrics such as factor-weighted returns, quantile analysis, and Sharpe ratio will be employed. Portfolio optimization will be performed using the risk model and factors with various optimization formulations. Quotemedia's end-of-day data and sector data from Sharadar will be utilized.
Project 5: NLP on Financial Statements (generate Alpha Factors from 10-k) Utilizing NLP techniques on corporate filings like 10Q and 10K statements, this project involves cleaning data, text processing, and feature extraction. Company-specific sentiments will be generated using bag-of-words and TF-IDF methodologies. Based on these sentiments, investment decisions regarding optimal timing for buying or selling will be made, aiming to generate an alpha factor. Quotemedia's end-of-day data and Loughran-McDonald sentiment word lists will be utilized.
Project 6: Sentiment Analysis with Neural Networks (LSTM) This project focuses on building a deep learning model to classify sentiment in messages sourced from StockTwits, a social network for investors and traders. LSTM networks will be constructed and trained for sentiment classification, with subsequent application to news data for signal generation.
Project 7: Combining Signals for Enhancing Alphas (using Machine Learning) Here, you'll combine signals using a random forest approach to enhance alpha. Addressing the challenge of overlapping samples, you'll integrate data from Quotemedia's end-of-day and sector data from Sharadar.
Project 8: Backtesting (Barra data) This project involves constructing a realistic backtester utilizing Barra data. The backtester will optimize portfolios with transaction costs while prioritizing computational efficiency for reasonably fast backtesting. Additionally, performance attribution will be conducted to identify the primary drivers of portfolio profit-and-loss (PnL). Customization options for the backtest will also be explored.