What an Engineer-Turned-Investor's Career Says About Building Responsible AI Governance

Most people who shape the rules governing new technology never touched the technology itself. That gap matters more than it used to. As artificial intelligence systems move from research labs into hospitals, courtrooms, and power grids, the people writing the guardrails need to understand not just the risks on paper, but how the machinery actually works underneath.

Eileen Tanghal's career traces that bridge in real time. She started as an electrical engineer, the kind of training that teaches you to think in circuits and constraints, in what a system can and cannot do under stress. She later moved into venture capital, where the job shifts from building technology to deciding which technologies deserve resources, trust, and a path to market. That transition, from engineer to investor, mirrors a broader shift happening across the AI policy world right now: technical literacy is becoming a prerequisite for good governance, not a nice-to-have.

Think of it like the difference between a building inspector who has actually poured concrete and one who has only read the code manual. Both can tell you whether a structure meets the standard. Only one of them understands where the standard might be wrong, or where it hasn't caught up to a new construction method. AI regulation is full of that second kind of gap right now: rules written by people who understand the stakes but not always the substrate.

The Case for Technical Fluency in Policy Roles

Tanghal's journey illustrates why organizations like IEEE, the world's largest professional body for engineers, have leaned so heavily into standards work for AI and related technologies. IEEE Standards exists precisely because someone has to translate abstract principles, safety, fairness, transparency, into technical specifications that a company's engineering team can actually build against. Without that translation layer, good intentions in a policy document stay just that: intentions.

This is the quiet, unglamorous work that rarely makes headlines but ends up mattering enormously. A regulation that says an AI system must be "explainable" sounds reasonable until an engineering team has to decide what threshold of explainability satisfies a court, a regulator, or a patient. Someone with hands-on technical experience is far more likely to spot when a requirement is vague to the point of being unenforceable, or so rigid it locks out beneficial innovation.

Venture capital adds another layer to this picture. Investors like Tanghal sit at a chokepoint in the innovation pipeline. They decide which startups get funded, which product ideas get a runway, and, increasingly, which governance commitments get baked into a company's culture before it scales. A venture capitalist with an engineering background is better positioned to ask hard technical questions during due diligence: does this AI system's training data raise privacy concerns, does the model's failure mode pose real-world harm, does the founding team actually understand the limitations of what they've built. Those questions, asked early and by someone credible, can shape industry norms long before any regulator gets involved.

That's not a small thing. Formal AI regulation, at least in the United States, has moved slowly and unevenly compared to the pace of deployment. In the absence of comprehensive federal rules, a lot of the practical governance of AI happens informally, through investor expectations, corporate guidelines, and voluntary industry standards. People with both technical depth and capital influence, exactly the profile Tanghal represents, end up filling gaps that legislation hasn't caught up to yet.

There's a risk in this arrangement worth naming honestly. When governance depends heavily on individual judgment rather than binding law, it becomes inconsistent by design. One investor's ethical bar isn't another's. A startup that gets funded by a values-driven VC might build responsibly; a similar startup funded elsewhere might not. That unevenness is exactly why formal standards bodies and regulatory frameworks still matter, even as individual leaders do important work inside that vacuum.

There's also a workforce dimension here that deserves attention. If the future of AI governance depends on people who can move fluidly between deep technical expertise and policy or investment decision-making, then the pipeline producing those people matters. Engineering programs, business schools, and professional societies all have a role in cultivating that hybrid skill set rather than treating technical training and governance training as separate tracks that never intersect.

Why It Matters

The stakes here reach far beyond any one career path or any one investor's portfolio. AI systems are already influencing decisions about who gets a loan, who gets flagged by a hiring algorithm, and how medical diagnoses get triaged. Get the governance wrong, whether through poorly designed regulation, standards nobody can actually implement, or investment decisions made without technical scrutiny, and the harm lands on ordinary people who never had a say in how these systems were built.

Careers like Tanghal's suggest a more hopeful path forward: one where the people making high-stakes decisions about AI's development actually understand its mechanics well enough to ask the right questions before problems reach the public. That doesn't replace the need for enforceable law or binding technical standards. But it does show what capable, technically grounded stewardship can look like while those slower systems catch up. The engineers who move into policy and investment roles aren't just changing careers. They're carrying essential fluency into rooms that badly need it.