Data Science Scientist

Henry Leung Yee Hin, ASA, MSc

Email: **************@*****.***; Mobile1: +65-86477419 (only when I am in SG); Mobile2: +44-740******* (UK); WhatsApp: +65-864****** / +44-740*******; CAREER SUMMARY

■ Career evolvements were driven by my ambition to take more sophisticated role.

■ Skills (maths, computer, finance) were proven to be portable across different job functions.

■ Path: Sales Data Operations Investment / Risk Quant & Data Science. SKILLS HIGHLIGHTS

Domain: Data Science, Sports/Betting, Risk Management, Quantitative Finance, Actuarial Science/Insurance. Technical: Deep Learning, Feature Engineering, Quantitative Methods, Cryptography, Unsupervised Learning. Management: Team Management, Project Management, Project Framing & Pitching. PROFESSIONAL AFFILIATIONS

Society of Actuaries (SOA)

■ Associate of the Society of Actuaries (ASA) Aug 2010 attained CFA Institute

■ Level 2 passed. Aug 2010 attained

LANGUAGES

English: Excellent; Cantonese: Native; Mandarin: Excellent Japanese: JLPT – N4 passed Jul 2023 attained

COLLEGE EDUCATION

The Hong Kong Baptist University

MSc in Applied Accounting and Finance. Sept 2009 – Apr 2011 The Hong Kong University of Science & Technology

BEng in Computer Engineering. Sept 2000 – Aug 2003 WORK EXPERIENCE

GreekIsGood Pte. Ltd.

Founder Jun 2022 –

Purposes, Tasks

■ Implement the Data Science & Quant ideas that are too controversial for other companies.

■ Leverage the developed models to manage my own investment portfolio.

■ Create an environment for me to work on after immigration, to fill the gap before receiving an offer.

■ Storage the IPs, and as a model/toy/art display cabinet. Great Eastern Life, Singapore

Data & Strategic Transformation – VP Principal Data Scientist Nov 2021 – Jun 2022 Summary

Assisted the company in the journey of migrating from on-premise to Cloud (AWS). A capstone project was initiated as the 1st attempt in the Cloud environment, the project’s needs have been the major driver of the migration plan.

- DS fundamentals: planned / created relevant tools in Cloud for the project as well as for future concerns.

- Architecture: suggested what the Cloud architectural components should be included.

- Connectivity: defined APIs and sync-up relevant parties for production.

- Projects review: reviewed existing projects to prepare migration in the future. Tokopedia, Singapore

Senior Principal Data Scientist Sep 2019 – Nov 2021 Lead Tasks

■ Framed projects for various teams, thus far the Recom, Risk, and Promo teams.

■ Led DSs to work on the projects and R&D tasks.

■ Led Risk Management personnel and DSs to transform to quantitative risk. Team Based Projects

■ Recom Team: Product Category Clustering.

■ Risk Team: Suspicious Shops Detection, Transaction Fraud Detection, Quantitative Loss Evaluation.

■ Promo Team: Merchant Voucher Optimisation.

Tools

■ Language: Python, Shell Script

■ Cloud Platforms: GCP: DataFlow, BigQuery, AI Platform, AWS.

■ Deep Learning: Tensorflow, Keras; Distributed: Apache Beam; DB: PostgreSQL. APJ Innovation Center, Hewlett Packard Enterprise, Singapore Data Scientist Lead / Manager Oct 2017 – Aug 2019

Senior Data Scientist Aug 2016 – Sept 2017

Lead Tasks

■ Managed team resources, headcounts, etc.

■ Delegated tasks and duties, coached members.

■ Organized team activities, e.g., standups, knowledge sharing, events, etc.

■ Appraised team members’ performances.

Development Tasks

■ Performed data ETL and sampling for text analysis projects

- Techniques: API calls, cron-job, error detection.

- Tools: Python, MySQL, Shell Script.

■ Built a sentiment measuring method on top of various NLP tools

- Techniques: linear regression.

- Tools: CoreNLP, TextBlob, NLTK, Python, Java, Shell Script.

■ Invented a sub-topic detection algorithm from a larger topic.

- Techniques: PCA, k-means clustering, path searching.

- Tools: Python, Spark.

- *Please check the Appendix at the end of the document.

■ Developed a filtering algorithm to filter out trash from the topic detection results

- Techniques: feature extraction, logistic regression.

- Tools: Python.

■ Rationalized highly unstructured, incomplete, obsolete, and various encodings data (Ongoing)

- Tools: Jupyter, Python, R.

■ Developed a system for a certain type of transportation optimization

- Techniques: convolution, linear regression, simulation.

- Tools: Tensorflow, Jupyter, Python, R.

■ Implemented an CNN model for predictive maintenance for a telco

- Techniques: CNN, ARIMA, EMA, STL Decomposition, PCA, etc.

- Tools: Tensorflow, Python, Tableau.

■ Invented an anomaly detection algorithm for a manufacturing solution startup

- Techniques: Fourier transform, probability normalization.

- Tools: Python.

Business Tasks

■ Crafted out business proposals, tenders, etc.

■ Evaluated timeline and resource.

■ Negotiated terms and conditions with clients.

■ Sought for projects funding.

■ Hosted workshops to prepare clients for PoC / projects engagements.

■ Visited clients for sales meetings.

Research Tasks

■ Clustering by applying the Demand & Supply concept

- Techniques: sentence similarity measure, PCA, demand & supply rules, logistic regression, greedy tree.

■ Theory and implementation of root cause analysis under different machine learning frameworks.

- Techniques: backward propagation and various mathematical tools.

■ Text files processing using unsupervised deep learning.

- Techniques: RNN, clustering.

Nugit, Singapore

Data Scientist Jul 2015 – May 2016

TAS II, Hong Kong

Researcher - Modeling Team Jul 2014 – Jul 2015

Summary

Employ StatArb (statistical arbitrage) strategy to bet on horse racing. The strategy is composed of 3 main functional areas:

- Data: Use raw, transformed, and derived data to quantify race information.

- Model: Train our models to reach the best estimates of parameters.

- Backtest: Generate probabilities based on the trained parameters and mimic real bets. The modelling team was formed by PhD, MPhil, and professional designation holders like me, from background of Stats/Maths/A.I./ActSci. We are responsible for model development, as well as to initiate the 2 other areas’ works. We know the input and backtesting requirements well, thus we are also highly involved in prototype development of factors (data) and betting strategies (backtest).

Tasks

■ Developed and modified machine learning models.

- Techniques: collaborative filtering, gradient descent, neural networks, boosting and bagging, PCA, etc.

- Tools: R, C++, SQL.

■ Built a platform that digests and process factor codes/classes created by various programmers.

- Techniques: parallel computing, OOP in Python.

- Tools: Python.

■ Created factors as inputs for the machine learning models.

- Techniques: clustering, Dijkstra's algorithm, clustering, t-test, Chi-square test, correlation test, etc.

- Tools: SAS, R, Python, Matlab, C++, SQL.

■ Devised investment performance analysis and prediction tools.

- Techniques: truncate distribution, convolution, goodness-of-fit test, etc.

- Tools: R, Matlab.

■ Automated backtesting and coped with the constraints (amount, advantage threshold, etc).

- Techniques: brute-force approach.

- Tools: shell script, Java, R, VBA.

■ Performed scenario analysis for the cases in-doubt.

- Techniques: mainly matrix manipulation.

- Tools: R.

Wing Hang Bank, Hong Kong

Middle Office Senior Analyst May 2013 – Jul 2014

■ Developed an in-house R program to calculate equity positions’ Delta Gamma VaR.

■ Re-engineered and validated the the VaR of FX and IR treasury positions (Misys).

■ Validated pricing models for bonds, bonds with options, equity derivatives, etc.

■ Configured treasury system (Misys) to fit the needs of front and back office.

■ Automated data collection and distribution procedures. Hong Kong Zhongcai Finance Investment, Hong Kong

Equity Research Analyst Apr 2011 – Jul 2012.

■ Developed a market trend quantifying model.

■ Developed a portfolio optimization model to mimic the market.

■ Developed an A&H statistical arbitrage strategy to enhance portfolio performance.

■ Traveled to the US for one month to study the trend of the flooring industry. CAI Business Indepth, Hong Kong

Assistant Data Analyst Mar 2010 – Nov 2010

■ Automated information ETL (extract, transform, load) process.

■ Developed analytic software for equities, commodities, debts, etc.

■ Ensured data integrity and enhanced quality of data. Banks and Financial Institutions, Hong Kong

Relationship Manager / Training Officer May 2005 – Feb 2010 Sirius International (HK) Ltd.

Software Engineer Oct 2003 – Apr 2005

■ Designed and implemented Display Content Management and Inventory Control Systems. COMPETITIONS

Dextra (by invitation, private)

Reversible Data Protection Challenge Jan 2016 - Mar 2016

- One of the Top 5 teams, I joined as a single person team.

Contact this candidate