• Developed a new trading strategy in the US Equities market by researching international market signals.
• Enhanced forecasts of options volatility and created a model that surpassed implied market prediction’s through analyzing and researching different predictors that occur earlier in the day.

• Developed a full-stack project using Route53, Cloudfront, API Gateway, S3, Lambda, and DynamoDB AWS services.
• Created a tool that decreased incoming support tickets by 90% and support response wait times by 95%.
• Coded using React, Typescript, and Java to create CDKs, front-end website, and AWS lambda functions.

Built an AI neural network NLP translator used by over 5,000 people

• Planned, designed, and followed through on procedures to create a web crawler that visited over 6 million relevant websites and a high-quality Interlingua translator with Python, BeautifulSoup, PyTorch, and CUDA.
• Created the Interlingua Corpus Project, a corpus of over 1.2 million quality-controlled Interlingua sentences (a 4000%+ increase to the second largest Interlingua corpus) by analyzing Interlingua sentences and developing a web crawler. Also collected 80,000+ parallel Interlingua-English sentences and other useful data.
• Built and optimized the first ever public and open source Online Interlingua-English Translator, which scored a 42.54 BLEU score (representing a translator with high quality translations according to Google) by designing and implementing techniques such as normalization and proper noun replacement.
• Learned to develop websites in React; use git, Jupyter, Google Colab, remote servers; deploy Python on Heroku.

Visit the Interlingua Corpus
Try out the Interlingua-English Translator

• Developed programs in MATLAB and Python to analyze and visualize the growth and motion of tumors in mice over time.
• Created an application that reduced the daily post-scan analysis time of optical imaging data from 1 hour to 2-3 minutes, greatly increasing the efficiency of data analysis for the lab researchers.
• Built and analyzed over 50 visual 3D models using MATLAB to highlight where tumors exist inside a mouse's body.

• Worked on a time series, forecasting, and NLP summarization project with Arcus Project to deteect and explain outlier events.
• Sourced projects from companies for ML@B members to gain industry experience.

• Coordinated and managed the Berkeley Trading Competition that hosted students from all over the country.

• Researched topics with Prof. Dawn Song related to safety, security, and privacy in reinforcement learning and deep learning alongside adversarial robustness, anomaly detection, and data augmentation using Python and PyTorch.
• Collaborated to create an educational course that teaches machine learning safety

• Studied topics in NLP, computer vision, and reinforcement learning. Researched and presented how GPT-3 works.
• Optimized a RNN machine learning translator by implementing techniques such as word tokenizing and unknown word prediction, increasing the BLEU score of the language translator by 50%.

The club's purpose is to introduce schoolmates to the joys of computer science through collaborative coding projects using the Python programming language. I teach club members using self-made lesson plans, design and plan semester-long projects such as a Connect-Four-playing AI bot, and manage mini-hackathons.