Daniel J. Rodriguez, PhD, is a mechanical engineer, strategist, and technology leader whose career spans multiphase flow measurement, thermal-fluids engineering, business strategy, applied data science, and AI infrastructure.
Over more than 2 decades, he has worked across research, industrial measurement, energy systems, enterprise technology, and AI-enabled infrastructure while maintaining a strong engineering foundation rooted in systems thinking, validation, reliability, and continuous learning.
Rodriguez holds a PhD in mechanical engineering from the University of Wisconsin–Madison and currently serves as thermal architect at Dell Technologies.
In this interview, The Way Ahead explores how engineers can expand beyond their original specialties without losing their technical identity, how engineering principles transfer across industries, and why deep domain expertise remains essential in an increasingly AI-enabled world.
Ali Abu Ikhzam (AI): We really appreciate you taking the time to share your journey with us. You built deep expertise in mechanical engineering, experimental systems, and flow measurement before expanding into AI infrastructure and enterprise technology. What gave you the confidence, and perhaps the courage, to grow beyond a domain you already knew so deeply?
Daniel J. Rodriguez (DR): The motivation began with advice from a supervisor who, before moving into marketing, warned that my focus on multiphase flowmeter (MPFM) technology could become a career 'pigeonhole.' MPFM is a narrow niche within upstream oil and gas—rich in complexity and innovation but limited in scope.
I set out to broaden my impact across adjacent production domains such as artificial lift and managed pressure drilling, while leveraging my multiphase expertise. I proposed a new, cross–product-line technical leadership role to align with this trajectory. In parallel, I developed skills in coding, data science, and machine learning (ML) and I worked with career coaches specializing in tech leadership. That guidance expanded my perspective and made a transition into tech both feasible and compelling.
My former employer ultimately offered a role closely aligned with my proposal—a technical fellow position with cross-functional leadership responsibilities. However, shortly beforehand, a longtime mentor and leading expert in cooling solutions for IT and AI infrastructure encouraged me to pursue a thermal architect role at a Fortune Global 100 company, amid the early stages of the 2024 artificial intelligence (AI) investment surge.
With support from my mentor and coaches, I applied and received an offer. I chose to make a diagonal move from product-line management in oil and gas to an individual contributor role in AI and high-performance computing infrastructure. The shift involved a new industry, expanded thermo-fluids focus, and relocation. It was a deliberate step: trading management responsibilities for deep technical immersion to successfully execute the transition.
AI: During your early research work, including particle tracking velocimetry and experimental flow systems, did you already feel drawn toward complex systems integration and data-driven engineering problems?
DR: That early research built a strong foundation in experimental design, data acquisition, image processing, and multiphase flow. I learned that high-quality data is essential for innovation—simulations like computational fluid dynamics and multiphysics models must be grounded in experimental results. As complexity increases, nondimensional parameters become critical for extracting general trends from limited datasets. The same principle applies to ML and AI. Effective models depend on meaningful feature extraction from reliable and traceable data.
Initially, I worked on gas/liquid two-phase systems. Later, with MPFMs, complexity increased significantly. Like the classical three-body problem, adding water to hydrocarbon liquid and gas introduces nonlinear interactions that are difficult to predict. I often reminded my team that this difficulty created opportunities for learning, innovation, and differentiation.
Working in domains without clear standards or established solutions requires comfort with uncertainty. You must accept what you don’t know and use it as a driver for curiosity and problem-solving.
Farhan Khan (FK): Your internship experience at Oxy exposed you not only to field operations, but also to automation implementation in real industrial environments. How important was that early exposure in shaping your later interest in large-scale engineering systems?
DR: The most important lesson from that internship was that true expertise requires firsthand, boots-on-the-ground experience. It forces you to think early about how a solution fits into real operations and exposes the realities of long hours in harsh environments. I’ve seen many technically sound designs fail because they were developed with a limited mindset. There is no substitute for getting broad experience and perspectives—engineers can learn a great deal from skilled technicians, machinists, and manufacturing teams.
I also gained a strong appreciation for safety. Eliminating risk at the design stage delivers the greatest payoff. Administrative controls and PPE are essential, but they are secondary measures when risks aren’t engineered out early.
Today, as I work on risk mitigation in large data centers, I see direct parallels with the oil field. Many lessons transfer well—often in ways that are not immediately obvious in this emerging domain.
AI: What did multiphase flow measurement teach you as an engineer that continues to influence how you approach technical problems today?
DR: MPFMs taught me that not all hurdles are technical. Measuring in-line without separation challenged decades of operational reliance on test and production separators. The main hurdle was convincing end users that MPFMs were not just an alternative to separation, but that they enabled deeper insights into well behavior beyond traditional well testing and production allocation. The technology had matured well before users were ready to change their mindset. Sometimes the challenge is cultural—a people problem, not a technical one.
The second lesson is that even the best MPFMs (including traditional separators) will yield poor results without close collaboration with the operator. If inaccurate or outdated fluid thermophysical properties (PVT) data is used in flow calculations, measurement accuracy will suffer.
You must educate end users about the technology—often requiring nondisclosure agreements to share deeper knowledge. Just as I spent time educating MPFM users about the technology, I am now learning and teaching liquid-cooling concepts so that users accustomed to air cooling for decades can have a better understanding of what’s possible so that we can collaborate and explore new solutions.
FK: Many engineers view flow measurement and AI infrastructure as completely separate industries. What engineering principles do you personally see connecting these domains?
DR: AI infrastructure became a focus once direct liquid cooling of IT racks became a requirement to operate the latest generations of CPUs and GPUs. The flow rates required can be comparable to those seen in upstream production facilities. Also, you need to optimize cooling capacity and allocate heat load from individual IT racks to facility-scale heat capture and ultimate rejection to the atmosphere. Redundant, accurate, and repeatable flow measurements are needed to operate your facilities with high availability of the IT capacity while minimizing energy and make-up water utilization.
Some equipment is not yet constrained by industry standards. It is still the case with MPFMs, and it is even more pronounced with heat capture technologies like cold plates and coolant distribution units (CDUs). Innovations like the Dell PowerCool CDU C7000 raise the bar for cooling capacity for extreme-density AI racks.
In upstream production we think upwards from measurements at the well level, the pad level, into production facilities and ultimately to LACT units for fiscal-grade production measurements. Data centers are also arranged in a tiered fashion from individual servers to server racks to clusters of racks in the “white space” of the datacenter and then adding the large power and heat rejection infrastructure in “gray-space” of the data center. You can use your oilfield systems mapping as a proxy to the one used for hyper scale IT infrastructure.
AI: Do you think engineers ever truly leave their original engineering identity behind, or do they carry it forward in the way they frame problems, test assumptions, and judge reliability?
DR: Engineering is fundamentally about solving problems through a systematic, data-driven approach that accounts for practical constraints. It addresses uncertainty using both quantitative and qualitative methods. Engineers develop expertise in recognizing real-world implementation challenges that extend beyond purely technical concerns, including budget and schedule constraints.
With experience, engineers become adept at identifying opportunities for execution improvement and competitive differentiation. The core engineering mindset does not fade—it sharpens over time. What does fade over time is the narrow affinity to their core academic training domain (e.g., mechanical, electrical, chemical, civil), as practical experience broadens across disciplines.
Over time, engineers develop a shared vocabulary and transferable skill set. Experienced engineers recognize patterns and extrapolate effectively from prior work. They also excel in pre-mortem analysis—the ability to anticipate failure modes and assess their potential impact early in the design or implementation process. Many engineers successfully follow a managerial career development path or become successful entrepreneurs. Their core skills are an excellent platform beyond technical management. They thrive beyond their core technical skills if their soft skills are also developed in parallel with substantial business and financial acumen.
AI: Over the years, your education expanded beyond traditional mechanical engineering into business strategy, leadership, applied data science, finance, and AI. Was this broadening intentional from early on, or did it emerge gradually as your career evolved?
DR: That continuous learning path was very deliberate. Growing into an effective engineering manager and technical leader requires developing soft skills, along with business and financial acumen as I mentioned earlier. The data science and ML component began as a personal interest, yet it quickly became essential for further differentiation and for adding value to existing solutions. In particular, ML inference solutions deployed at the edge, where compute and networking are very significant constraints remains a challenge—it remains a relevant issue adjacent to my current role in the tech infrastructure industry,
You must be intentional and strategic in architecting your career path. Periodically defining your ideal role should be treated as an ongoing exercise—akin to continuous integration and continuous CI/CD in the software domain. This can be framed as identifying the areas in which you can grow most effectively to deliver sustained value while avoiding burnout or stagnation.
Similarly, as a manager, I encouraged my team members to proactively expand beyond their core technical skill sets. This often led to a step change in confidence and overall performance, both for individual contributors and for the team.
FK: You pursued certifications ranging from ML and statistical inference to strategy, negotiation, SQL, AI infrastructure, and organizational leadership. How do you decide which new skills are genuinely worth investing time in as an experienced engineer?
DR: I maintain the scout mindset. Looking out for emerging trends and pathfinding for technologies, methodologies, and frameworks that resonate with my personal interests and with the challenges, gaps, and opportunities that constantly evolve within the industry.
Another means of narrowing down the scope of continuous learning is taking the point of view of the end user. The end user is your customer, your service crew, a potential customer in a different segment. End users often implement and adopt gap-filling practices that you should have built into your solution. Let the end user’s workday inform what needs to be in your solution/product roadmap. The roadmap in turn informs what skills you need to deliver an effective solution within the budget and time constraints that stakeholders expect.
AI: For experienced engineers, learning something new can sometimes feel like starting over. How did you maintain your technical confidence while entering unfamiliar areas such as AI infrastructure, enterprise systems, and strategy?
DR: Curiosity, coupled with clear direction, drives the expansion of technical knowledge. Adopting a humble, apprentice mindset—one that encourages asking fundamental questions—is essential. Today, even AI tools, when prompted to act as subject-matter experts and cite authoritative sources, can accelerate early learning. Networking also plays a critical role. I had colleagues in oil and gas who began their AI/ML learning journeys well before the acceleration seen around 2022. We formed an informal learning group to share insights and resources. At least two members transitioned into data science and machine learning careers before 2022, positioning themselves to capitalize on emerging opportunities.
My study of strategy, particularly product strategy, highlighted how disruption often arises at the intersection of adjacent disciplines, guided by a beginner’s mindset. A key influence was the Jobs To Be Done (JTBD) framework, popularized by Clayton Christensen through his work on disruptive innovation at Harvard University and further elaborated by Mohanbir Sawhney at Northwestern University. The central idea is that a product is “hired” to perform a specific job.
Successfully addressing a JTBD often requires capabilities beyond your current skill set. The process involves identifying the job your target customer needs done, then mapping the gaps in your skills required to deliver a minimum viable product (MVP). That MVP becomes a focal point for directing curiosity and sustained learning.
FK: Many young professionals struggle with balancing technical depth against broad interdisciplinary learning. Did you ever worry that expanding into strategy, business, and AI-related fields might dilute your technical specialization?
DR: My experience was quite the opposite. My technical mindset enabled me to identify and study business case studies that were directly relevant to my customers, rather than being limited to more generic, abstract, or service-industry-oriented examples.
Accordingly, I reiterate that engineering provides a strong foundation for product management and offers a distinct head start in learning finance, accounting, and business analytics. I observed that many non-engineering participants in business and product strategy programs struggled with methods such as multivariable linear regression and conjoint analysis. For engineers, these are foundational analytical tools, often requiring little more than adaptation in terminology, variables, and constraints.
That said, soft skills are essential for developing meaningful business acumen. These are typically cultivated through direct engagement with customers, stakeholders, and partners. The stereotype of the introverted, narrowly focused engineer who dives too deeply into technical detail during such interactions can hold true for some individuals. However, it remains a stereotype.
Early-career professionals should actively work to overcome this perception by deliberately developing their interpersonal and communication skills. This requires patience, humility, and a disciplined focus on listening more than speaking. These interactions are not the appropriate setting to overwhelm others with technical depth, but rather an opportunity to build alignment, trust, and shared understanding.
AI: A lot of public discussion around AI focuses on software and models. Your work frequently highlights architecture, deployment, cooling systems, hardware integration, and operational scalability. Do you think the future of AI will increasingly depend on traditional engineering disciplines?
DR: Traditional engineering disciplines will remain essential. Data centers depend on the optimal design of power and cooling infrastructure, which ultimately drive their key performance metrics. AI can already enhance an individual engineer’s productivity by mapping their unique skills and experience into AI agents for consistent, repeatable applications. All without requiring the skill of a professional developer to implement it. With a coding assistant using today’s thinking models you can create useful agents by describing your problem specification and solution methods in plain English. You still need to validate what is generated. Human expertise and experience remain in the loop. Andrej Karpathy’s recent car wash example where AI can still advise that you walk instead of drive to the car wash highlights that LLMs lack the context that only you with your expertise and common sense can provide.
From my point of view, software and models are not the primary constraints at this stage for scaling-up AI infrastructure. Instead, real estate and power availability—along with capacity, efficiency, and reliability in power and cooling systems—are the dominant limiting factors.
FK: Your recent work around AI factories, hybrid AI deployment, liquid cooling, and enterprise infrastructure suggests that AI systems are becoming deeply physical systems, not just digital ones. How important will thermal management, energy systems, and infrastructure engineering become in the next phase of AI growth?
DR: I addressed this at a high level in my prior response, but I can provide additional context here. The trend is toward higher-power silicon, lower case temperatures, and potentially increasing overall or localized heat flux. Air cooling is not going away; however, leading-edge platforms increasingly require direct liquid cooling in combination with air cooling. Some platforms are already expected to transition to fully liquid-cooled architecture.
Rising heat flux may also drive greater focus on developing practical and environmentally responsible phase-change cooling solutions.
If two-phase flow becomes a required capability, my technical path may come full circle, as I began my work studying two-phase flow phenomena at the scale of hundreds of microns. There is a plausible scenario in which I could pivot back to addressing the complexities of multiphase flow in the context of next-generation compute platforms—potentially including multiphase flow measurement at the data center or even individual server level.
This could become a Job To Be Done for AI infrastructure within the next few years.
AI: In several recent LinkedIn posts, you highlighted the importance of integration across hardware, software, cooling, deployment, and operational ecosystems. Has systems integration become the defining engineering challenge of modern AI infrastructure?
DR: Yes. AI doesn't stress one layer of the stack; it stresses all of them simultaneously.
In traditional environments, teams could optimize infrastructure component by component. AI changes that. A denser GPU cluster shifts cooling requirements. Cooling design affects rack layout and power distribution. Network architecture influences model performance. Every decision cascades, which means the engineering challenge is no longer building powerful systems—it's making the entire system work as one.
Forums like the Open Compute Project describe this moment as a transition from ‘building the plane while flying it’ to a more mature state, one characterized by fully operational digital twins of the facility, its systems and associated equipment. Systems engineering is what enables that shift, moving organizations from reactive integration to intentional design.
The competitive advantage in modern AI infrastructure isn't better components. It's better orchestration across the full system.
FK: In your recent geothermal commentary, you emphasized operational feasibility, thermal expansion management, and peer-reviewed validation despite the excitement surrounding the technology. Has industrial field engineering shaped how you evaluate emerging technologies today?
DR: Emerging technologies inevitably experience growing pains, particularly when they are conceived and designed far from their intended operating environment. At one extreme, consider a Mars rover, where accurately simulating the operating environment is an explicit project requirement. At the other extreme is something like a CAD workstation, where the design and operating environments are effectively identical.
For the many scenarios that fall between these extremes, conducting a site survey remains the best practice. Even after operating within dozens of seemingly similar environments, there are always potential pitfalls when deploying a design that may appear physically sound and robust under controlled conditions. The phrase “field-proven” is commonly used in marketing materials for a good reason.
AI: In today's technology landscape, there is often enormous pressure toward hype cycles, especially around AI. How important are engineering skepticism and validation culture when evaluating transformative technologies?
DR: Frankly, the pace of development in today’s technology landscape—particularly since the emergence of AI and its rapid permeation across domains—makes it difficult to keep up with new developments and to distinguish between hype and meaningful progress.
We will likely need to leverage AI tools to summarize and distill relevant information. At the same time, we should not rely on AI to serve as the final arbiter of the viability or significance of that information. Maintaining a healthy level of skepticism is essential—not only toward new claims, but also toward AI as a validation tool, especially in areas where models may lack up to date training data or domain depth to effectively assess the underlying assumptions.
AI: Your career combines technical depth, leadership, strategy, and emerging technology domains. Did your transition into leadership roles change the way you think about engineering problems themselves?
DR: One of my career coaches frequently emphasizes that technical teams are rarely hindered by the inherent complexity of the problems they are solving. More often, initiatives lose momentum—or fail altogether—due to people-related challenges. As a leader of technical teams, your soft skills are critical in identifying where those obstacles exist.
People need clear guidance on objectives, along with an understanding of how their roles—and those of their teammates—contribute to overall success. At the same time, when you hire top talent, they expect autonomy to make meaningful decisions. They should not fear pushbacks from leadership; rather, they should expect support, particularly when navigating setbacks or failures.
Failing to address people-related challenges will ultimately undermine an organization’s ability to attract and retain top talent. Effective leaders trust their teams to solve technical problems while actively removing non-technical barriers within workflows. This requires stepping back, delegating effectively, and focusing on engineering the team rather than performing the engineering work on the team’s behalf.
Mutual trust is the foundation for addressing people-related issues. Building and maintaining that trust depends on active listening in one-on-one interactions, as well as consistent clarity and transparency in communication with the broader team.
FK: Many young professionals today feel pressured to choose between staying deeply technical, moving into leadership, or adapting to AI-enabled industries. Your path seems to combine all three. Did that evolution happen by design, or did it emerge over time?
DR: My advice is to avoid viewing this choice as binary, discrete, or permanent. Over the course of my career, I have moved from an individual contributor role to technical management and product management, and then back to an individual contributor role in a new industry and domain—all driven by a deliberate desire to grow and learn.
Early in one’s career, it is important to build strong technical depth to establish credibility as a leader. At the same time, leadership is not defined by a title. It is situational and can be exercised at any level. As a frontline contributor, you can demonstrate leadership when you are given the autonomy to make decisions and act in a manner consistent with your experience and track record.
It is essential to be intentional about accepting new challenges and leading projects—particularly those that push you beyond your comfort zone. Equally important is the ability to decline opportunities that are not aligned with your broader goals, thereby preserving your time and energy. Be deliberate in architecting your career, and do not fixate on titles. If you are growing with clarity and purpose, leadership will emerge as a natural outcome.
It is also worth noting that successful organizations increasingly recognize that higher compensation is not exclusively tied to management roles. Their success depends on the expertise and commitment of deeply technical contributors. This is evident in the ongoing competition for top technical talent among big tech companies.
AI: Looking back, was there a specific period when you realized your career trajectory was expanding from technical specialization toward broader systems-level leadership and strategy?
DR: I don’t believe there was a single defining period or clear inflection point. In my case, it has been more of a series of situational role shifts rather than formal changes in title or job description. We often say—and leadership frequently expects—that we wear many hats, sometimes even within a single day. Over time, these situational roles have a cumulative effect. Today, I feel better prepared to lead teams and define strategy as a frontline contributor than I did when those responsibilities were explicitly tied to my title and role.
I view this evolution as analogous to the shift in software engineering from a Waterfall model to Agile and CI/CD practices. It is not a linear progression, but rather a series of iterative cycles of growth and expansion. Expecting a linear career path with clearly defined milestones is, frankly, an outdated mindset—perhaps inherited from the more rigid, hierarchical model of military rank structures with strictly defined chains of command.
While that model may have been effective in earlier decades, organizations that continue to rely on it today risk falling behind more agile, adaptive competitors. In the current environment, flexibility, continuous learning, and the ability to navigate evolving roles are far more critical to sustainable success.
FK: As AI infrastructure, energy systems, and industrial technologies continue converging, where do you think engineers from traditional industrial disciplines may have an advantage compared with purely software-focused professionals?
DR: A data center can be viewed as a token-processing facility, where token throughput ultimately drives revenue. Margin, in turn, depends on optimizing the inputs and processes that enable that throughput. Power consumption typically ranges from tens to hundreds of megawatts per facility, requiring strong backup systems and redundancy.
Cooling infrastructure is evolving rapidly to support liquid-cooled architectures. Water and coolant recirculation systems are scaling from approximately 4-in to 8-in piping to handle increased heat capture. The associated mechanical systems must also be designed with redundancy in mind. Captured heat must either be rejected to the atmosphere or repurposed, for example, through district heating systems.
Power supply options are diversifying and may include on-site generation via gas turbines, photovoltaic solar arrays, geothermal sources, and potentially small modular nuclear reactors. This gray space infrastructure underpins the IT equipment in the white space, and it is increasingly the limiting factor in system performance and scalability.
Right now hardware is very top of mind—not only in terms of IT components, but also in the supporting infrastructure. This reality continues to drive demand for innovation among electrical and mechanical engineers, particularly in improving system efficiency, capacity, and reliability.
All this infrastructure must be in place before any tokens can be processed. By comparison, software can appear almost secondary, often split between open-source and proprietary ecosystems. Most engineers today possess at least some degree of software proficiency, and AI-driven coding assistants are increasingly democratizing development, enabling engineers to build specialized applications and AI agents.
Highly specialized roles—such as field-programmable gate array developers—will continue to be extremely valuable, although the total number of such positions may decline over time. Meanwhile, there is significant opportunity for a new generation of engineers in traditional disciplines, whose skills will need to be augmented by AI tools just to keep pace with the rapid rate of change across domains.
AI: What technical or interdisciplinary skills should younger engineers begin developing today if they want to remain relevant in the next decade of industrial and AI-driven transformation?
DR: Disciplines such as ethics, philosophy, and bioscience will become increasingly important to ensure that safety remains paramount, intellectual property is protected, and technological disruption results in a sustainable step change in global well-being across diverse populations.
AI: Finally, after moving across research, industrial measurement, energy systems, enterprise strategy, and AI infrastructure, what has remained most constant in your engineering mindset throughout the journey?
DR: For most of my career, I would have said curiosity and the habit of going back to basics i.e. asking, ‘Why does this actually work?’ rather than just accepting the answer. Those instincts have served me well across very different industries.
However, AI is changing the pace of everything, and I think we need to add something: the ability to keep learning, even when what you learned last year may already need updating. In fast-moving fields like energy and AI, knowledge has a shorter shelf life than it used to. That's not a reason to panic. It's a reason to stay curious and humble.
What AI tools do extraordinarily well is process information at a scale and speed no human can match. What AI tools don't have is judgment, the ability to decide what matters, or to weigh the unintended consequences of a decision. That part is still on us. And I think that's a good thing.
It's why I believe fields like ethics and life sciences will become more important, not less, as AI becomes more capable. Technology is a tool—how we direct it and what we choose to optimize for reflects our values as people and as an industry.
Ali Abu Ikhzam, SPE, is a service engineer at TechnipFMC, specializing in multiphase meters. His work focuses on projects and services for Metering Systems, with interests in project management and business development. He is the young professionals (YPs) representative on the SPE Flow Measurement Technical Section Board, a member of the SPE Stavanger Section Young Professionals Board, and the founder and YP chair of the SPE Libya Section. He has been an active SPE member for over 10 years, contributing to technical and professional development initiatives across the Society. He holds a master's degree in petroleum engineering from Universiti Teknologi Petronas and is passionate about digitalization, automation, and energy transition.
Farhan Khan, SPE, is a field engineer at Weatherford, specializing in artificial lift systems and digital solutions through the company’s Next Gen Leadership Development Program. His work focuses on drilling and production technologies, digital solutions, and field operations, with interests in leadership development and energy transition. He is the Young Professionals representative on the SPE Flow Measurement Technical Section Board, program chair for the Australasia of Drilling Systems Advancement Technical Section, president and communication chair of the IADC Australasia Young Professionals Subcommittee, and a director and Young Professionals chair of the SPE Northeast India Section. He has actively contributed to professional development initiatives across both SPE and IADC, fostering collaboration and engagement among young professionals. He holds a master’s degree of professional engineering in petroleum engineering from Curtin University, Australia, and is passionate about innovation, digital transformation, and developing the next generation of energy professionals.