Applying industry - proven knowledge and practices of data science, data-driven prognostics, data mining, and machine learning for real world problems. More than twenty years of working experience and domain knowledge in development and implementation of quantitative modeling statistical modeling, and risk management for energy and technology industries. AWS Cloud/Elastic Search/R/Python/Matlab/SAS/SQL/C/C++/ working skills. Technical/Team leadership.
Director of Data Science
- Mentoring and building a dynamic Data Science team, currently 3 members
- Working in the AWS Cloud based Elastic Search/Kibana Platform to support data engineering team for Data Ingestion, Data Quality Control, Business Alerts with machine learning and Python scripts
- Serve as Energy Quantitative Modeling and Data Science Expert and Consultant to build strategic and sustainable business process for Software development, deployment and moving to production
- Developing pricing, optimization, risk management solutions for oil products' scheduling, trading activities
Senior Data Scientist
- As Lead Data Scientist for Global Trade Screening projects, designed, developed and implementing Natural Language Processing based Python coded intelligent and automated screening tools to replace 80% of manual screening process.
- As Lead Data Scientist for Enterprise Direct Cash Forecasting projects, designed developed and implementing Machine-Learning forecasting methodology combined with driver based statistical models.
Senior Data Scientist
- Working at Corporate Business Intelligence Division, providing Predictive
- Analytics modeling, support, and implementation for Executives, Business Departments such as Operations, Sales, Marketing, Call Centers with cutting edge Data Mining, Machine Learning, and Data Analytics.
- Designed, implemented and executed marketing and sales Campaigns for Customer Retentions and Renewals
- Built Customer Churn Model for customer behaviors with Decision Tree/Deep Learning/Survival Analysis methodologies
- Established Customer Channel Comparison models with data mining processes
- Developed Customer Risk and Value Scores with advanced data analytics
- Implemented Customer Renewal Model with predictive analytics
Data Scientist Manager
- Led the design, development, and delivery of DATA tool (Datalake Analytics Access Tool). It utilized Confidential ’s newly developed Hadoop BIG DATA infrastructure, retrieving high volume data in terabytes from HDFS based Proficiency Database with automatic data quality checking and preprocessing, generating tables and downloading data into desktop in the backend with HAWQ capability. It is 150 times faster comparing to traditional data analytics access tool.
- Developed and implemented algorithms in Matlab to detect abnormal behavior and operation of power engines in real time.
- Built Bayesian algorithms in R to estimate failure rate for Gas Turbine products: LM 6000, LM 2500, as well as LMS 100 using past ten years failure and testing data.
- Developed Python based Random Forest algorithm to conduct importance factors analysis for Weibull modeling of engine failure.
- Invited TED Talk to Confidential Analytics Team: Hadoop and its Applications
- Developed forward price curve models of US Natural gas and basis, WTI Oil, as well as NGLs for evaluation of Natural Gas reserves and assets transaction deals.
Risk Modeling Consultant
- Developed new methodologies for statistical and mathematical multivariate survival models to combine financial and physical risk for oil drilling process by using MATLAB, VBA, Minitab, and Java as platform.
- Conducted daily risk predictions for Middle East, US Permian, as well as European Continent.
- Used Monte Carlo simulation in the model prediction process to provide customers realistic presentation of the risk distribution based on the model parameter precision.
- Conducted statistical clustering analysis for data analysis and model accuracy improvement using Minitab and MATLAB
- Developed new model process and methodological framework to improve the multivariate survival modeling process with Weibull, Lognormal and Linear distribution.
- Coded Monte Carlo simulation in Java to automate the current risk prediction process.
Risk Analytics (QAD) and Business Analyst
- Perform risk and quantitative model validation and design for Confidential ’s Aligne product.
- Generate and provide official documentations to demonstrate, illustrate, and validate step by step mathematical computation results, business process as well as data flow for Aligne Analytics product’s energy risk management applications, such as price simulation, principal component analysis, PDEs computation, Monte Carlo VaR calculation, stress testing, option models, and power generation models, and so on.
- Serve as liaison in product management between client business and development team, to provide business requirements/system function requirements documentations, including screen design, dataflow, and computation process for trading system development implementations.
- Provide consultant service to Confidential ’s sale team, professional service team, as well as customer service team in terms of energy risk management control and processes, as well as quantitative modeling expertise.
- Provide quantitative and bushiness function testing design to replicate customer’s issues/bugs related to Aligne Analytics and Trading product.
- Provide system testing environment for the entire Aligne Analytics and Trading product.
- Design and conduct periodic quantitative and functional validation and regression testing for energy risk management core computations, such as profit & loss, cash flow, Mark to Market value, option modeling, and so on.
- Work side by side with Confidential ’s Development Team to debug and solve the issues related to Aligne Analytics and Trading product.
Independent Risk Consultant
- Provide detailed documentations for current mathematical algorithms, system processes, database data structures as well as information integrating flow with OpenLink and RMS system.
- Propose and specify the next generation of Stress Testing systems’ requirements, methodologies, processes as well as system functions and integrations with OpenLink and RMS systems.
- Conducted model review and validation for the current Full Revaluation Stress Testing Tools for natural Gas, Gas Skew, Oil, Oil Skew, US Power, Base Metal portfolios
- Provide overview, high level, mid-level, as well as low level technical documentation of the existing system’s methodologies, processes, Access database data structures, and data flow interaction with OpenLink and RMS system.
- Propose and specify the new integrated Stress Testing Systems for all commodities to overcome the scalability problems the current Stress Testing Tools encountered.
Independent Risk Consultant
- Built pilot Excel/VBA based models to capture the overall Client Company’s financial and physical business and risk management process by utilizing all the available corporate data
- Developed Monte Carlo simulation based financial and physical risk models for the Client Company to analyze the risk distribution and statistical characteristics for the future 20 years revenue generation and forecasts.
- Evaluate and calculate the financial benefits and revenue impacts of implementing the future proposed financial/physical risk management systems for the Client Company.
Lead Risk Specialist
- Participated and implemented a Monte Carlo based cross-commodities market risk management system for Confidential . Developed and implemented a Matlab based off-system VaR model for BP’s Canadian Gas portfolio to validate and improve the functions and capabilities of Confidential ’s Entegrate Analytics System. Develop a Matlab based off-system VaR model for BP’s NGL and Power portfolio to validate and improve the functions and capabilities of Confidential ’s Epsilon System.
- Conducted model validation and review for the mathematical methodology in the Confidential ’s Long Term Risk Management Framework. This system uses a spot price simulation processes based on mean-reverting methodology to evaluate the value and risk of the long term Gas and Power transactions.
- Participated in design, development, and validation/calibration of various deal/transaction/option evaluation models and systems for Power and NGL trading desk; including compute/assess market/credit risk for various risk factors in these models such as market fundamental changes, forward market movements, volatility as well as correlations.
- Daily Var and risk metrics report for NGL trading Book for senior management, provide front office daily VaR impact analysis of potential new deals; participate in weekly TSPA meeting to provide trading manager and commercial manager risk analysis and assessment of major trading strategies, provide statistical analysis on Confidential ’s portfolio actual P&L and theoretical P&L distribution, stress testing, liquidity computation, as well scenario analyses.
- Participated in and validated the simulation methodology and parameter estimation for BP’s Power Credit Risk PFE model, involved in the development and implementation of BP’s brand-new SAS based cross-commodities Credit Risk Management system including PFE methodology, which uses the historical forward market price movement’s seasonal and monthly correlation matrix to simulate and compute the potential future exposure for the power industry.
- Led and coordinated development efforts between Risk function and IT function for Confidential Risk management systems in architect design, data flow integration, and control process, as well as methodologies.