Personal details
Prashant J.
Timezone: Melbourne (UTC+10)
Summary
Expert computer software engineer with 15+ years of experience in developing complex multithreaded applications in C/C++ on Unix platforms
Strong experience in Python(numpy, scikit, pandas, keras etc) specifically using it for Data Science projects involving machine learning. Certified deep learning specialist from Coursera.
Expertise in code debugging and interest in learning new technologies.
Experienced with PostGreSQL, SQLite, code versioning tools, bug tracking tools
Basic knowledge of Go and Java
Personal Projects
2019
Pandas
Machine Learning
Supervised Learning
Reinforcement Learning
Unsupervised Learning
PyTorch
Numerai is a data science company that powers the Numerai hedge fund. They provide encrypted stock market data that has 310 features and is distributed across different time-eras.
The training data comprises of 120 eras with each era having ~4000 test points. The validation data is 12 eras. The test set has 54 eras and then there is live data which is the actual stock market data on which the predictions are made.
I have used supervised and unsupervised algorithms, pipelines, Auto-encoders etc concepts to come up with my models where one of the model was ranked as high as #4 in the leaderboard