Orion Darley

I am responsible for creating new data science-based solutions and leading the development from scoping, development, and operationalization on client-facing analytics teams.

Serve as the lead data scientist for two major Oil & Gas company's Analytics COEs.

Additional capacities where I deliver: manage data scientist direct reports, develop thought leadership on various data science-related topics, partner with vendors on technical architecture opportunities, facilitate monthly meetings for a data science region (~150 data scientists), support opportunities/contract staffing, am chief of staff for regional AI project contract reviews, and develop responses to RFPs (<$15M).

• Developed and implemented an machine learning operations (MLOps) framework (deploy, execute, monitoring, reporting and feedback) for a major O&G client Analytics COE; includes managing multiple deployed MVPs, facilitating workshops on platform integration capabilities, analytics portfolio prioritization, documentation standards, and developing governance.
• Direct response to the U.S. House Committee on Science, Space and Technology: Designed and operationalized signal condition monitoring logic to identify, detect, and classify SCADA time series signals where methane gas is directly released into the atmosphere for 1500 O&G wells; integrated into enterprise surveillance system, detects a ~100 events daily where events have been manually detected at ~180 per year.
  
• Created and implemented a two-year advanced analytics roadmap for an O&G analytics COE specific to integrating new and existing technology, adopting new data science methodologies, data integration, and upskilling in data science.
• Enhanced optimization model for a client which provided decisions on what path of wells multiple maintenance vehicles should take to minimize lost production, minimize cost of service, and maximize NPV per well; integrated into SOP. 
• Advisor/Presenter to a client's data literacy program for 3 years (~350 engineers). Coached 15 engineering teams on applying data analytics and machine learning to domain specific engineering problems; won a corporate global excellence award (finalist #4 of 100), I won the "Top Instructor and Mentor" Award.

I lead the overall delivery of a 13-person, $4.2M engagement with an O&G analytics COE where we delivered and operationalized analytics solutions to business use cases. My overall responsibility was to deliver planned margin within the boundaries of the solution plan while also managing any risks, issues and deviations to the Analytics COE in the course of delivery.

I was responsible for creating new data science-based solutions and leading the development from scoping, development, and operationalization on client-facing analytics teams.

Served as the lead data scientist for three client's Analytics COEs.

• Designed and operationalized signal processing (conditional monitoring) and dashboard to identify, detect, and classify SCADA time series signals to alert potential oil well equipment failures; operationalized this dashboard, integrated into enterprise surveillance system, est. NPV savings of $4.8M.
• Designed and operationalized recursive multistep, ensemble deep learning model and dashboard to predict market pricing; received positive feedback from CEO and C-level; Advisor to Chief Trading Officer.
• Designed and operationalized multivariate survival models and dashboard to create a risk scoring system and ops dashboard(s) to support maintenance prioritization for $100M in annual maintenance spend; Monitored for 18 months for performance drift, coefficient changes, data leakage, regressions in estimation.
• Developed propensity models used by sales team to support prioritization of existing target customers based on CRM factors for a client’s customer intelligence organization; received positive feedback from CIO.
• Designed and operationalized production-ready, data-driven models to forecast well production to support estimation of the economic value for installing several different types of well pumps; operationalized with dashboard.
• Designed and operationalized a reporting automation and event detection solution that identifies subsea equipment patterns and sends email alerts to user groups for specific conditions.
• Lead prioritization of 40+ analytics initiatives and design of KPI frameworks for 50+ marketing programs for a technical architecture modernization project.
• Supported an initiative focusing on generalizing ML algorithms used to predict engine failures on two Deepwater oil rigs; integrated into Plant Operations Advisor (POA), the largest cloud-based program used to prevent unplanned downtime on large oil rigs. 
• Supported an initiative focusing on predicting equipment failures (ML/Survival) used for oil well treatments; Featured at Spark Summit 2020.

Led a three-person ML team, developing and implementing NLP models for various applications, including text classification and Named Entity Recognition used to identify specific information in clinical charts and categorize chart pages by type (i.e. labs, orders, encounters, DOB, etc.).

• Developed an AI Clinical Chart Review Solution, subjects including: text preprocessing for NLP, ML performance evaluation, and ML monitoring.
• Developed a Named Entity Recognition (NER) AWS pipeline using AWS Comprehend (NLP), SageMaker, Glue, Athena, and Quicksight for classification and visualization of 10,000's business article entities; AWS now offers this as a similar solution, “AWS Hot Topics”.
• Served as an analytics advisor to support high-level data and analytics technical architecture, and development of an analytics roadmaps.

I lead a start-up data science and engineering team that developed machine learning-based solutions on operations/supply chain data for enterprise analytics tools. I was also responsible for scoping, developing, and delivering data science use cases for HPE Global Operations.

• Supported the development of the Global Operations Analytics Control Tower responsible for all analytics roadmaps and prioritization in organization. 
• Lead deployment of several forecasting models to predict price movements of high-value commodities for the server memory market, supporting a high volume (M+) purchasing strategy for a enterprise app; one-of-few tools used to power negotiating purchase pricing.
• Lead design and operationalization of ML algorithms to predict the expected sales order “Wait Time” and “Delivery to First Commit”-related metrics from order creation to manufacturing, prior to availability of all components.

I provided strategy consulting for all global business units, leadership and support teams within HPE Servers (≈$15B). I provided insights into server-market investment decisions and identified opportunities and risks through the use of strategy roadmaps, analytics enablement, competitive analysis and executive forums such as strategic programs, strategy development workshops, and global business unit-level requested analysis.

I developed predictive models and research for the Net Promoter Score (NPS) System. Analyses and problem-solving work included determining priority drivers of NPS and customer loyalty surveys, researching regional revenue impact of NPS, and evaluating touch point gaps in customer surveys.

DCI is a consulting firm that specializes in leadership training for the offshore O&G business. I was responsible for developing analytics for client assets, F2F marketing, and CRM management. Partners include Ensco, Technip, Diamond Offshore, and several others.

I diagnosed and corrected malfunctions for both major and component-level assemblies on the F-15 and F-16 fighter aircraft using schematics and wiring diagrams, integrated test systems or other measurement devices for various equipment including: attack control, instrument, flight control, communications, navigation and others.

(2011-2013) Competed under the World Class Athlete Program (WCAP)