[Online] Hong Kong Machine Learning Meetup Season 5 Episode 2

![[Online] Hong Kong Machine Learning Meetup Season 5 Episode 2](https://secure.meetupstatic.com/photos/event/9/e/c/8/highres_508720648.webp?w=750)
Details
This month's meetup will be online, we have 3 great Kaggle Competitors that will share with us their journey on the platform and some of the projects that they've submitted in the past!
We will be hosting an in-person event with Refinitiv in DEC - stay tuned!
1st Speaker - Rajneesh Tiwari
Rajneesh leads the Product and Strategy teams at Bulian AI. Before founding Bulian AI, he built multiple machine learning systems at Novartis, Ericsson, and various research-focused startups in India. He had more than a decade worth of experience in building ML systems delivering high impact for customers. Along with pursuing his MS from Georgia Tech, he is also very active on Kaggle and is a practising Kaggle Master (top 1%) with multiple Kaggle victories under his belt.
Presentation Introduction:
Session Presentation Topic: Rajneesh will present his approach to an online AI competition about detecting and predicting presence of life-supporting chemical compounds on Mars with the help of data collected by NASA’s Mars rovers. This competition was hosted on DrivenData and Rajneesh scored 8th position on the private Leaderboard.
2nd Speaker - Caleb Yung
Caleb is currently at HSBC as a data scientist focusing on using machine learning to detect financial crime from transactions. In his free time, he enjoys learning AI/ML in competitive environments and is a Kaggle Competition Expert with several competitions ending up in the top 5%.
Presentation Introduction:
Caleb will present his solution to the Optiver Realized Volatility Prediction competition, which asked for a good ML solution to predict the US stocks’ 10-mins window realized volatility using the past 10-mins order book data. The submitted models went through a 3-month live validation and were evaluated by RMSPE (Root Mean Squared Percentage Error). Caleb will walk through his overall modelling pipeline (boosting models, deep learning, and meta models) and some creative time series feature engineering ideas that led to his final top 5% ranking.
3rd Speaker - Patrick Yam
Patrick Yam worked as a quantitative researcher in a Hedge fund, focusing on solving challenging problems using machine learning. He is a Kaggle competition master (Top 100 on the global competition leaderboard) with 4 gold medals in various Kaggle competitions.
Presentation introduction:
In the G-Research Crypto Forecasting competition, we used our machine learning expertise to forecast short-term returns in 14 popular cryptocurrencies. We are given a dataset of millions of rows of high-frequency market data dating back to 2018 which we can use to build our model. after the submission deadline has passed, the final score is calculated over the following 3 months using live crypto data as it is collected.
In this competition, I built a deep learning model which only receives raw data as the input, and features engineering is almost not required. The final ranking is 7th place out of ~2000 teams.

[Online] Hong Kong Machine Learning Meetup Season 5 Episode 2