Back to Insights

The PE Department Advisory Board Pitch: Why Every Program Needs an AI/Data Science Track

Dr. Mehrdad Shirangi | | Published by Groundwork Analytics LLC
Editorial disclosure

This article reflects the independent analysis and professional opinion of the author, informed by published enrollment data, industry hiring trends, and direct experience with both petroleum engineering education and AI deployment in upstream operations. The author holds a PhD from Stanford in optimization applied to energy systems and has worked with operators, service companies, and technology providers since 2018. The proposed curriculum reflects professional judgment, not a commercial offering.

If you sit on a petroleum engineering department advisory board, you have watched enrollment numbers decline for a decade. You have heard students ask whether petroleum engineering is a dying field. You have seen talented undergraduates choose computer science, data science, or mechanical engineering instead. And you have probably noticed that the graduates who do finish PE programs are getting hired immediately -- often before they walk across the stage -- because the industry's demand for technically trained engineers has not gone away even as the supply has collapsed.

This article is a pitch. Not for a product, but for a curriculum change. The argument is straightforward: petroleum engineering programs that add a structured AI and data science track will reverse enrollment declines, produce graduates who are more valuable on day one, and strengthen the industry partnerships that keep departments funded.

The data supports this. The industry is hiring for it. Several programs are already doing it. And for advisory board members who want to champion this change, the talking points and funding sources are laid out below.

The Enrollment Crisis Is Real, But the Job Market Is Not the Problem

PE enrollment in the United States has dropped approximately 75% from its 2014 peak. The number of ABET-accredited petroleum engineering programs has been halved, from roughly 35 to 20. Some programs that survived did so by merging with other departments, broadening their scope to "energy engineering" or "subsurface engineering," and hoping that the rebrand would attract students who would have otherwise walked past the department.

These are alarming numbers. But they obscure an important fact: the job market for petroleum engineers remains exceptionally strong. Over 90% of PE seniors receive job offers before graduation. Petroleum engineering consistently ranks as the highest-paid engineering discipline for new graduates, with the National Association of Colleges and Employers reporting a median starting salary of $100,750. The industry is not shrinking its workforce -- it is struggling to fill positions because fewer students are entering the pipeline.

The enrollment decline is not driven by lack of jobs. It is driven by perception. High school students and their parents associate petroleum engineering with climate risk, boom-bust cycles, and an industry that may not exist in 30 years. Whether those perceptions are accurate is beside the point. They are real, and they are keeping students away.

Curriculum modernization will not single-handedly fix a perception problem. But it can meaningfully help in two ways. First, students who are interested in AI, data science, and automation -- which is to say, a large and growing share of engineering-minded students -- will see a PE program with a data science track as more relevant and forward-looking than one without it. Second, industry partners who are desperate for graduates with both domain expertise and computational skills will increase their engagement with programs that produce them. More scholarships, more internships, more capstone project sponsorships, more reason for students to enroll.

What Industry Is Actually Hiring For

The disconnect between what PE programs teach and what operators need has been widening for years, but the data is now stark enough to demand attention.

The Hiring Data

AI-related job postings in oil and gas have increased 78% year over year, while the qualified talent pool has grown only 24%. This gap is not closing. It is widening. The average age of an oil and gas worker in the United States is 56, and only 12% of the industry workforce is under 30. Every company planning an AI or digital transformation initiative -- and most are -- faces the same fundamental constraint: there are not enough people who understand both the domain and the technology.

A survey by the Energy Workforce & Technology Council found that 45% of upstream companies offer zero AI or data science training to their technical staff. Only 15% of reservoir engineers report frequently using machine learning in their work. These are not companies that have decided AI is not valuable. These are companies that have not been able to hire or train people who can implement it.

What the Job Postings Say

A review of recent upstream technology job postings reveals a consistent pattern. Companies are looking for engineers who can:

  • Work with production data at scale. SCADA systems, historians, time-series databases, and the messy reality of field data -- missing values, sensor drift, unit inconsistencies, and the fact that most production data arrives via a daily email from a pumper.
  • Build and deploy machine learning models for operational problems. Decline curve analysis, anomaly detection on artificial lift systems, well performance prediction, drilling parameter optimization, and automated surveillance.
  • Integrate AI into existing workflows. This is not research. Companies want engineers who can take a trained model and embed it into a production monitoring dashboard, connect it to a SCADA feed, and make it useful to a field engineer who does not know what a neural network is.
  • Write production-quality code. Python is the baseline. SQL is expected. Familiarity with cloud platforms (AWS, Azure, GCP), version control (Git), and data engineering tools (Airflow, dbt) is increasingly requested.

The common thread: these roles require petroleum engineering domain knowledge combined with computational skills that most PE programs do not currently teach. A mechanical engineer or computer scientist can learn Python, but they cannot intuitively grasp why a 200-well rod pump field generates anomaly detection challenges that look nothing like a manufacturing quality control problem. Domain knowledge matters. But domain knowledge without computational fluency is increasingly insufficient.

Programs That Are Already Doing It

Several petroleum engineering programs have recognized this gap and are actively building AI and data science into their curricula. Their approaches vary, but they share a common recognition: the PE degree of 2030 must look different from the PE degree of 2010.

Texas A&M University

Texas A&M's Harold Vance Department of Petroleum Engineering offers a Graduate Certificate in Data Analytics for the Petroleum Industry, a 13-credit-hour program that can be completed alongside a master's degree or as a standalone certificate. The certificate includes courses in statistical learning, data-driven reservoir modeling, and machine learning applications in petroleum engineering.

What makes Texas A&M's approach notable is its integration with industry. The department's advisory board includes data science and digital transformation leaders from major operators, and the certificate program was explicitly designed to address the skills gap that those industry partners identified. Students completing the certificate graduate with both PE domain expertise and a credential that signals computational competence to employers.

Colorado School of Mines

The Colorado School of Mines offers a Graduate Certificate in Petroleum Data Analytics, designed for both current students and working professionals. The program focuses on the intersection of data science methods and petroleum engineering applications, with coursework covering statistical methods, machine learning, and data management for subsurface applications.

Mines has also invested in its broader data science infrastructure, with the university-wide Center for Data-Driven Discovery supporting cross-disciplinary research that includes energy applications. PE students at Mines have access to data science coursework, computing resources, and research opportunities that would have been unavailable a decade ago.

University of Oklahoma

OU has taken an ambitious approach: offering a dual master's degree in Petroleum Engineering and Data Science. This is not a certificate bolted onto an existing degree -- it is a fully integrated program that produces graduates with graduate-level credentials in both fields.

The dual degree requires additional coursework and time, but it produces graduates who are genuinely bilingual in petroleum engineering and data science. For students willing to invest the extra semesters, the career payoff is substantial. These graduates enter the workforce with a rare combination of skills that commands premium compensation and opens doors to roles in digital transformation, AI deployment, and technology leadership.

University of Wyoming

The University of Wyoming recently renamed its petroleum engineering department chair position to focus on "Subsurface Energy and Digital Innovation" -- a signal that digital technology is not an add-on to the department's mission but central to it. Wyoming's approach reflects the reality that for smaller programs competing for students and faculty, differentiation through digital and AI capabilities may be a survival strategy.

Wyoming's model is worth watching because it represents what is possible for mid-sized programs that cannot afford to build an entirely separate data science curriculum. Instead, they are integrating digital and AI concepts into their existing PE courses and leveraging university-wide data science resources to fill gaps.

A Proposed Curriculum Addition: The 4-Course AI/Data Science Track

For departments that want to move beyond one-off electives, here is a concrete proposal: a 4-course (12-credit) AI and Data Science track that can be offered within an existing PE program. This track is designed to be additive -- it does not replace core PE coursework in reservoir engineering, drilling, production, or petrophysics. It supplements it.

The courses are sequenced to build skills progressively, from foundational programming through applied machine learning to emerging technologies. Each course uses petroleum engineering examples and datasets so that students see the connection between computational methods and the problems they will face in the field.

Course 1: Python for Petroleum Engineering

Prerequisites: None beyond standard PE prerequisites (calculus, physics) Semester: Junior year, Fall

This is the foundational course. It teaches Python programming through petroleum engineering examples, so that students are writing code to solve problems they recognize from their other coursework.

Topics include:

  • Python fundamentals: variables, data structures, functions, control flow
  • NumPy and Pandas for engineering data manipulation
  • Matplotlib and Plotly for data visualization
  • Reading and writing common petroleum data formats (LAS files, CSV exports from production databases, WITSML data)
  • Basic statistics and data cleaning -- handling missing values, unit conversions, outlier detection
  • Introduction to version control with Git

Capstone assignment: Given a dataset of 500 wells with production history, write a Python script that loads the data, cleans it, generates decline curves for each well, and produces a summary report with visualizations.

Why this comes first: Students cannot do machine learning, data engineering, or automation if they cannot write code. And they will not learn to write code in a generic computer science class with the same engagement they will find when the examples involve drilling parameters, decline curves, and well logs. PE-specific programming courses have higher completion rates and better learning outcomes because students see immediate relevance.

Course 2: Data Engineering for E&P

Prerequisites: Course 1 (Python for Petroleum Engineering) Semester: Junior year, Spring

The industry's single biggest barrier to AI adoption is not algorithms -- it is data. This course teaches students how petroleum data is collected, stored, managed, and prepared for analysis.

Topics include:

  • SCADA systems: architecture, data flow, common platforms (Emerson, ABB, Weatherford)
  • Time-series databases and historians (OSIsoft PI, InfluxDB)
  • Relational databases and SQL for production data
  • Well log data formats: LAS 2.0, LAS 3.0, DLIS
  • WITSML for real-time drilling data exchange
  • Data quality: systematic approaches to identifying and handling missing data, sensor calibration issues, and measurement errors
  • Data pipeline basics: extracting data from source systems, transforming it, and loading it into analytics-ready formats (ETL)
  • Cloud storage and computing fundamentals for petroleum data (AWS S3, Azure Blob Storage)

Capstone assignment: Build a data pipeline that ingests production data from a simulated SCADA system, stores it in a SQL database, runs quality checks (identifying wells with missing or anomalous readings), and produces a daily surveillance dashboard.

Why this matters: A machine learning model is only as good as the data it is trained on. Graduates who understand data engineering -- where petroleum data comes from, why it is messy, and how to clean it -- will be dramatically more effective than graduates who only know how to call scikit-learn functions. Most failed AI projects in oil and gas fail at the data layer, not the model layer.

Course 3: Machine Learning for Subsurface Applications

Prerequisites: Course 2 (Data Engineering for E&P) Semester: Senior year, Fall

This is where the domain expertise and computational skills converge. Students learn machine learning methods through petroleum engineering applications, focusing on the problems that operators are actually trying to solve.

Topics include:

  • Supervised learning: regression and classification with PE applications
  • Decline curve analysis using ML (Arps, stretched exponential, and ML-based approaches)
  • Well performance prediction from completion and geological parameters
  • Lithology classification from well log data
  • Unsupervised learning: clustering and dimensionality reduction
  • Well analog identification
  • Production pattern recognition
  • Reservoir characterization from seismic attributes
  • Anomaly detection for production surveillance
  • ESP failure prediction
  • Rod pump dysfunction identification
  • Gas lift valve performance monitoring
  • Time-series forecasting for production data
  • Model evaluation, validation, and the unique challenges of small-sample petroleum datasets
  • Responsible AI: understanding model limitations, avoiding overfit, and knowing when a physics-based model is better than a data-driven one

Capstone assignment: Using a real (anonymized) production dataset, build and evaluate a machine learning model that predicts which wells in a rod pump field are likely to experience failures in the next 30 days. Present results to a panel that includes industry practitioners.

Why this sequence: By this point, students can write code (Course 1), understand where petroleum data comes from and how to prepare it (Course 2), and are now ready to build models that solve real problems. The course is taught in the context of petroleum engineering applications -- not generic ML -- which means students understand both the technique and the domain nuances that determine whether a model will work in the field.

Course 4: AI Agents and Automation in Operations

Prerequisites: Course 3 (Machine Learning for Subsurface Applications) Semester: Senior year, Spring

This is the most forward-looking course in the track, and it addresses a technology shift that is already underway in the industry. Major service companies like SLB, Baker Hughes, and Cognite have deployed agentic AI systems -- software agents that can autonomously monitor operations, make recommendations, and take actions. This course prepares students for that reality.

Topics include:

  • Large language models (LLMs) and their application in petroleum engineering
  • Retrieval-augmented generation (RAG) for technical document search and Q&A
  • Agentic AI frameworks: autonomous agents that can observe, reason, and act
  • Model Context Protocol (MCP): how AI agents connect to data sources and tools
  • Hands-on work with open-source MCP servers (including petro-mcp, an MCP server designed for petroleum engineering data)
  • Building AI agents for operational workflows: automated surveillance, exception-based monitoring, report generation
  • Integration with existing enterprise systems (SCADA, production databases, ERP)
  • Ethics, safety, and human-in-the-loop design for autonomous systems in safety-critical operations
  • Industry case studies: SLB Tela, Baker Hughes Leucipa, Cognite Atlas AI

Capstone project: Design and build an AI agent that monitors a simulated field of 50 producing wells, detects anomalies, generates diagnostic reports, and recommends interventions -- with a human approval step before any action is taken. Students present to an industry panel.

Why this course exists: Agentic AI is not a theoretical concept in oil and gas. Equinor reported $130 million in AI-driven savings in 2025. Saudi Aramco documented $1.8 billion in value from AI initiatives in 2024. The companies deploying these systems need engineers who understand both the domain and the technology. Graduates who have built an AI agent for petroleum operations -- even a prototype -- will enter the workforce ahead of their peers.

Industry Advisory Board Talking Points

If you are an advisory board member who wants to champion an AI/data science track, here are the arguments that resonate with faculty committees, deans, and provosts.

For the Faculty Committee

  • This does not replace core PE. The 4-course track supplements the existing curriculum. Students still take reservoir engineering, drilling, production, petrophysics, and completions. They graduate as petroleum engineers who can also code, manage data, and build ML models.
  • Enrollment will benefit. Students choosing between PE and computer science -- a decision that many prospective students face -- are more likely to choose a PE program that includes a data science track. The track signals that the department is forward-looking, not anchored to a curriculum designed in the 1990s.
  • Over 90% of PE seniors get jobs before graduation. The enrollment problem is not about employability -- it is about perception. A modern curriculum helps fix the perception problem.
  • PE remains the highest-paid engineering major for new grads at $100,750 median starting salary. Combining that earning potential with AI/data science skills makes the value proposition even stronger.

For the Dean and Provost

  • AI job postings in oil and gas are up 78% year over year, while the qualified talent pool grew only 24%. This is a documented supply-demand gap that PE programs are uniquely positioned to fill.
  • Industry partners will fund it. Companies that cannot hire enough AI-capable petroleum engineers will sponsor scholarships, donate computing resources, fund faculty positions, and provide datasets for coursework. This is not speculative -- the programs already doing it (Texas A&M, Mines, OU) report strong industry engagement.
  • Cross-disciplinary collaboration is fundable. NSF, DOE, and SPE Foundation all have funding mechanisms for curriculum innovation at the intersection of energy and data science. A proposal that combines PE domain expertise with AI/data science education fits squarely within current funding priorities.
  • The department's survival may depend on it. Programs that do not modernize are the ones that will close next. The contraction from 35 to 20 programs is not over. Differentiation through AI and data science is a defensible strategy for remaining relevant.

For Industry Partners on the Board

  • The 45% problem: Nearly half of upstream companies provide zero AI/data science training. Programs that graduate AI-capable petroleum engineers directly reduce your training costs and time-to-productivity.
  • The age gap: With the average O&G worker at 56 and only 12% under 30, the industry faces a knowledge transfer crisis. AI-literate graduates can accelerate knowledge capture by building systems that encode domain expertise -- the digital twin of the retiring engineer's judgment.
  • Only 15% of reservoir engineers frequently use ML. This is not because ML does not work for reservoir problems. It is because most reservoir engineers were never taught how. Fix the curriculum and you fix this number over the next decade.
  • Your competitors are already investing. If your company is not engaging with PE programs on AI curriculum, your competitors are. The companies that build relationships with forward-thinking PE departments now will have first access to the best graduates.

Funding Sources: How to Pay for It

Curriculum additions require funding for faculty, computing resources, software licenses, and course development. Here are the most viable funding sources for an AI/data science track in PE.

Industry Sponsorship

Operators and service companies have a direct financial interest in producing AI-capable PE graduates. Concrete asks for industry partners include:

  • Named scholarship programs for students in the AI/data science track ($5,000-$15,000 per student per year)
  • Computing resource sponsorship -- cloud computing credits from AWS, Azure, or GCP education programs (often available at no cost through existing academic partnerships)
  • Dataset donations -- anonymized production, drilling, or reservoir datasets for use in coursework and capstone projects (operators have terabytes of data that could be anonymized and shared)
  • Adjunct faculty support -- funding for industry practitioners to co-teach courses, particularly Course 4 (AI Agents and Automation) where real-world deployment experience is essential
  • Capstone project sponsorship -- companies provide real problems, data, and mentorship; students provide solutions; both benefit

DOE and Federal Grants

The U.S. Department of Energy funds education and workforce development initiatives through multiple programs:

  • DOE Office of Fossil Energy and Carbon Management -- workforce development grants that align with energy technology education
  • NSF Improving Undergraduate STEM Education (IUSE) -- grants for curriculum innovation in engineering, including cross-disciplinary programs that combine domain expertise with computational skills
  • NSF Research Experiences for Undergraduates (REU) -- supplements that can fund student research at the intersection of petroleum engineering and data science

SPE Foundation

The Society of Petroleum Engineers Foundation funds scholarships, faculty development, and educational initiatives that support petroleum engineering education. A proposal for an AI/data science track aligns with the Foundation's stated priority of keeping PE education relevant to industry needs.

University Internal Funding

Most universities have internal mechanisms for funding curriculum innovation:

  • Provost's curriculum development grants -- typically $10,000-$50,000 for new course development
  • Cross-departmental collaboration funding -- grants that incentivize collaboration between PE, computer science, and data science departments
  • Online/hybrid course development funding -- many universities invested in online education infrastructure during and after COVID; an AI/data science track with online components may qualify

What Groundwork Analytics Offers

Groundwork Analytics works at the intersection of AI and petroleum engineering. We have built the tools, written the code, and deployed the models that this curriculum prepares students to use. For PE programs developing an AI/data science track, we offer the following:

Guest Lectures and Workshops

Dr. Mehrdad Shirangi (Stanford PhD, Optimization/Energy Systems) is available for guest lectures on AI applications in petroleum engineering, machine learning for subsurface problems, and the practical realities of deploying AI in upstream operations. These are not sales pitches -- they are technical presentations grounded in real project experience.

Capstone Project Mentorship

We mentor capstone teams working on AI and data science projects in petroleum engineering. This includes problem scoping, dataset preparation, technical guidance, and evaluation. We bring real industry problems to capstone projects, so that students work on challenges that operators actually face.

petro-mcp: An Open-Source Teaching Tool

We built and maintain petro-mcp, an open-source MCP (Model Context Protocol) server designed for petroleum engineering data. petro-mcp provides:

  • 9 tools for accessing and analyzing petroleum engineering data
  • 2 resources for structured data access
  • 5 prompts for common petroleum engineering analysis tasks
  • A fully tested, documented codebase that students can study, extend, and use in coursework

petro-mcp is particularly relevant for Course 4 (AI Agents and Automation in Operations). Students can use it as a foundation for building AI agents that interact with petroleum data -- learning MCP, agent frameworks, and petroleum data access in a single project. It is open-source, free, and designed to be extensible.

For faculty developing Course 4 or similar coursework, petro-mcp provides a ready-made platform that eliminates the need to build data access infrastructure from scratch. Students can focus on building agents and solving problems, not on plumbing.

Industry Context and Curriculum Advice

We track the upstream O&G software landscape, production operations technology, and AI adoption trends across the industry. For departments designing curriculum, we can provide current data on what tools and technologies operators are actually using, what skills they are hiring for, and where the biggest gaps exist between academic preparation and industry needs.

The Urgency: Why Now

Three trends make 2026 the right time for PE programs to act.

The AI adoption curve is steepening. Thirteen percent of oil and gas organizations have already deployed agentic AI systems, and 49% plan to begin deployment in 2026. The AI in oil and gas market is valued at $4.28 billion and growing at 13% CAGR. The industry is not experimenting with AI anymore -- it is scaling it. Graduates who enter the workforce in 2028 or 2029 need to be prepared for an industry where AI agents, automated surveillance, and ML-driven optimization are standard tools, not novelties.

The enrollment window is closing. Programs that have already contracted are harder to rebuild. Faculty lines, once lost, are difficult to recover. Industry partnerships, once severed, take years to re-establish. The programs that modernize now will attract the students and funding that sustain them. The programs that wait will face further contraction.

The workforce gap is structural. An average worker age of 56, only 12% under 30, and a generation of domain experts approaching retirement. The industry needs young engineers who can build the AI systems that capture, encode, and operationalize the knowledge that retiring experts carry in their heads. This is not a problem that solves itself. It requires intentional investment in education.

Conclusion: The Case Is Made. Now Make the Motion.

The data is clear. PE enrollment is down 75%, but the job market is strong. Industry needs AI-capable petroleum engineers, and it cannot find enough of them. Programs at Texas A&M, Colorado School of Mines, the University of Oklahoma, and the University of Wyoming are already building AI and data science into their curricula. A 4-course track is achievable within existing PE programs without displacing core coursework. Funding is available from industry partners, federal agencies, and SPE Foundation.

If you sit on a petroleum engineering department advisory board, the next step is straightforward: put this on the agenda. Propose a committee to evaluate the feasibility of an AI/data science track. Bring the enrollment data, the hiring data, and the examples from peer programs. Offer to connect the department with industry partners who will sponsor it.

The students who choose petroleum engineering in 2027 will graduate into an industry that runs on AI. Make sure they are ready.

Dr. Mehrdad Shirangi is the founder of Groundwork Analytics and holds a PhD from Stanford University in Energy Systems Optimization. He has been building AI solutions for the energy industry since 2018. Connect on X/Twitter and LinkedIn, or reach out at info@petropt.com.

Related Articles

Looking for O&G Jobs?

Petro-Jobs uses AI to match your resume to 79+ curated oil & gas positions.

Try Petro-Jobs

Have questions about this topic? Get in touch.

Back to Insights