Colin Sturm, Ph.D., P.E.

Professional Info

Digital transformation leader with 21+ years in oil and gas and a PhD, PE in Chemical Engineering. Architect of production-grade AI/ML solutions for upstream, midstream, and downstream operations including agentic AI, enterprise data platforms, and industrial optimization. Proven track record of translating complex petroleum engineering challenges into cloud-native solutions that deliver measurable business outcomes for Fortune 100 energy companies.

Work Experience

Amazon - Denver, CO

2020 - Present

Details

Senior Energy Consultant

• Subject matter expert in support of AWS Global Energy/Oil & Gas Accounts.
• BPX Energy: Agentic AI Operations Assistant (2024–2025)
• Led development of the first agentic AI operations assistant deployed to production in the energy sector, a voice-enabled digital assistant for field operators.
• Delivered 5% production uplift and ~50% operator productivity improvement; CEO public endorsement and Bloomberg media coverage.
• Occidental Petroleum: Drilling & Completions Digital Transformation (2024–Present)
• Technical lead on multi-year platform integrating 6+ vendor data pipelines, GenAI chatbot, and ML prediction models.
• Achieved 400% ROI and 72 hours of engineering time eliminated per workflow cycle (12,000 hrs/yr); presented at CERAWeek 2026.
• Baker Hughes: AI/ML Platform Strategy (2025–Present)
• Originated and grew a Fortune 500 oilfield services account from zero to four concurrent AI/ML workstreams within 8 months, establishing CIO-level strategic relationship.
• Designed platforms for subsea asset intelligence, tender response automation, and geoscience workflow modernization.
• Continental Resources: Supply Chain Transformation (2024–Present)
• Architected cloud-native supply chain platform reducing supplier onboarding from 45 days to 5 days and eliminating $50M/year in production delays.
• VP Supply Chain: "Colin worked as a direct extension of our team... brought valuable industry knowledge and deep technical expertise"
• GE Vernova: Enterprise AI Transformation (2026)
• Led technical strategy for AI program spanning 29 use cases targeting $4B+ in quality cost reduction.
• Designed multi-partner engagement model coordinating three SI partners; secured executive approval for dedicated architecture team.
• Shell: Subsurface Data Platform Leadership (2021–2026)
• Led 7 engagements deploying the energy industry's open subsurface data platform, the only global implementation on AWS at the time.
• Designed multi-region architecture serving 80,000+ users; pioneering work directly informed a commercial product offering (EDI).
• NextEra Energy: Capital Allocation Analytics (2022–2024)
• Built analytics platform affecting $900M/year in well development decisions, delivering 90% faster decision-making.
• Succeeded where a prior consulting firm failed; CEO and COO endorsements.
• Phillips 66: ML Refinery Process Optimization (2023–2025)
• Independently developed neural network models achieving 92.5% accuracy, outperforming incumbent vendor.
• Introduced SHAP explainability and dynamic what-if analysis for safety-critical refinery operations.
• Boardwalk Pipeline: Project Intelligence Platform (2026–Present)
• Designed and built AI-powered project health dashboard aggregating 4+ enterprise systems into unified executive view.
• Additional
• TotalEnergies (well integrity, France), Cenovus/Vital Energy (GenAI workover automation), AWS Energy Symposium (GenAI workshop, 90+ leaders).

Apache Corporation - Houston, TX

2014 - 2020

Details

Senior Production Engineer

• Dynamic Optimized Pumping Routes
• Developed an algorithm in Python to optimize pumper routes based on production, risk, and proximity.
• Reduced drive time and mileage by 30% and pumper's daily route automatically directed to 75% greater potential cash margin.
• Scaled out to all of Apache North America: running 15% of Apache NA routes, $250M/D, 50 MBBLD, 200 MMSCFD, with more than 100 users of the optimization algorithm.
• Created a mobile report with asset tracking and relevant well site information for pumpers on site.
• Project managed a team of 10 in multiple disciplines: ESRI mapping, field implementation, data management, programming, and statistics.
• Predicted seismic induced activity
• Collaborated with Auckland University researchers to develop an algorithm that predicts pressure buildup and local seismic activity of salt water disposal.
• Autonomous ISIP picking reducing human error, increasing confidence and accuracy.
• Coauthored best paper out of 600 submissions awarded by US Rock Mechanics Symposium.
• Completion data collection
• Created an algorithm to run on a Raspberry Pi, connect to a service company network, and transmit relevant data to Apache.
• Reduced third party companies from data flow lowering per well cost by $3000.
• Cut the time needed to bring data to near time from days to seconds.
• Lessened the potential for error by reducing the necessity for human interaction.
• Extended analysis of StimPlan simulations
• Rock fracture modelling of injected chemicals, surfactants, proppant, and water, with a team of 3 covering disciplines of reservoir, petrophysics, and geology.
• Launched a Python web framework accessible within Apache network to house multiple applications.
• Time of Stimplan analysis reduced from days to minutes using Python algorithm.
• Organized and led a monthly completion data networking and governance meeting for all North America.

Production Engineer III

• Optimized production wells in Oklahoma (SCOOP/STACK) and Texas (Austin Chalk).
• Tank leak detection
• Used machine learning in Python with Cygnet and OSI PI to improve leak detection of holding tanks. Reduced false positive +30%, and increased true positives 50%.
• Initial Shut-in Pressure escalation and plateau
• Developed a workflow in Spotfire utilizing Python and R to estimate fracture dimensions based on published theory.
• Frac AFE Cost Estimation
• Improved workflow regions used to estimate the cost of completing a well.
• Designed a well completion database used by all districts.
• Managed rotational engineers and interns.
• Created numerous data analysis charts using Spotfire.
• Developed decline curves from Aries data showing daily granularity.
• Created database and imported Frac Focus data for analysis.

Image Version Gif Version

• Quickly assess actual production to forecasted production.
• Filters to visualize problem areas.

• Choose any well or group of wells to compare production.

• Created entirely using Python and querying the Aries database.
• Select criteria for which wells to analyze.

• Select a specific well to view the daily production forecast.
• Calculate monthly production for a group of wells.

Devon Energy via M&H Engineering - Houston, TX

2013 - 2014

Details

Contract Engineer

• Developed a Valve Fail Open graphical user interface for Devon Energy.
• Designed intuitive interface for easy and efficient valve analysis.
• Managed project directly with Devon operations engineers for 30% under budget final product.

• Created as specified by field engineers at Devon.
• Numerical methods used to determine accurate valve fail open pressure.

Procter & Gamble (P&G) - Brussels, Belgium

2011 - 2013

Details

Process Development Engineer

• Optimized a manufacturing process to realize a costs savings of $2.5 million/year.
• Led design, validation, and implementation of a $2 million customized 3D finite element method (FEM) particle tracking software using linear programming (with Altair Engineering).
• Created, developed and led a global 5 day training course in Python programming. ppt
• Designed an internal global engineering software for pipe flow and rheology modelling, which saved P&G $500MM.
• Developed linear least squares regression for fast curve analysis.
• Improved plant design by creating and validating static mixer simulations.
• Engaged vendors on technical issues for development and improvement at plant sites.
• Provided expert-level guidance to plant sites and work process implementation.
• Prepared and maintained the associated technical content for all developed models and research.

• Quickly determine the pressure drop, friction factor, and flow rate of a pipeline.

• Easily find the best curve fit for rheological data.

Image Version Gif Version

• A viscous fluid (red) mixing with a low viscous fluid as the fluids travel axially through a static mixer.
• One of hundreds of CFD simulations showing two fluids mixing axially down a static mixer.

Image Version Gif Version

• One very viscous fluid (green) mixing in a low viscous fluid in a stirred tank.
• The two fluids are being mixed by the impeller.
• The impeller appears stationary in the simulation for reduced simulation time and quick analysis.

WinSim Incorporated - Houston, TX

2007 - 2010

Details

Process Simulation Engineer

• Maintained and developed process simulation, similar to AspenTech or Hysys.
• Modelled projects: compressors, pumps, separators, etc. in surface and offshore.
• Researched and programmed a packed column option which increased sales more than 30%.
• Organized a weekly meeting with peers to review technical questions and brainstorm solutions.
• Remediated problems by extensive research of thermodynamic and engineering theory.
• Solved simulation errors with more than 200 documented solved issues.
• Maintained and updated a 400+ page technical document describing improvements.
• Worked with three phase flow, PVT curves, various equations of state.

• Worked closely with CEO on improving simulation thermodynamics.

• Created concise documentation on all changes and improvements.
• Documentation written completely in LaTeX totalling more than 400 pages.

Education

Doctor of Philosophy in Chemical Engineering

2002 - 2007

Details

University of Florida, Gainesville, FL

• Dissertation: Thermodynamic nano-filament force generation.
• Developed Matlab and Fortran algorithm to calculate energy stored in filament networks.
• Invented a micro-fluidic flow cell from parts in the laboratory saving $10K in essential supplies.

• A Listeria monocytogenes with actin polymers under scanning electron microscopy (SEM).

• A 50 nm bead with actin polymers under transmission electron microscopy (TEM).

• A 50 nm bead with actin polymers under total internal reflection fluorescence microscopy (TIRFM).

Bachelor of Science in Chemical Engineering

1997 - 2002

Details

University of Oklahoma, Norman, OK

• Capstone Project: Natural gas transportation and optimization of delivery.
• Minor in Chemistry, Graduated with Distinction.
• Collaborated with graduate students in the development of hard sphere simulation.
• Optimized Fortran code to simulate hard sphere/surface interactions.
• Wrote Fortran code to analyze Ranque-Hilsch vortex tube temperatures for spot cooling.
• Created algorithm in a General Algebraic Modeling System (GAMS) for my Capstone Project. Excerpt

• An example schematic I created for the technical report from unit operations.

Publications

Technology

Spoken Languages

License and Certification

Texas Professional Engineer License #121960

Volunteer Work