Computer Science Degrees Are Not Obsolete — The Entry-Level Pipeline Is the Problem

This article is republished from the IEEE Spectrum careers newsletter, produced in partnership with Parsity, a tech career development company.

The Degree Isn't Dead — The Pipeline Is

A lot of people are telling newly graduated engineering students that their degree was a mistake and that AI will replace them before they land their first job. I disagree.

I have 12 years of experience in software engineering, have conducted more than 100 interviews on both sides of the table, and run Parsity, an AI engineering training program. In observing who actually breaks through in today's job market, a few consistent patterns have emerged. Here is why I believe the market is not as grim as it appears — and what I would do if I were looking for my first tech job.

The Data Needs Context

The Federal Reserve Bank of New York recently reported that the unemployment rate for recent U.S. computer science graduates stands at 6.1%, and 7.5% for computer engineering graduates. By comparison, philosophy graduates come in at 3.2% and art history graduates at 3.0%. Those figures look alarming at first glance, but most headlines leave out critical context.

When researchers factor in underemployment — graduates working in jobs that do not require a college degree — engineering-related fields perform relatively well, with underemployment rates below 20%, compared to an average of 42% across all recent graduates. Many fields reporting low unemployment rates do so because graduates have simply accepted jobs unrelated to their field of study. When unemployment, underemployment, and early-career earnings are considered together, computer science and computer engineering remain among the strongest-performing disciplines in the broader labor market.

The degree is not the problem — the hiring pipeline is. Job postings listed as "entry-level software engineer" grew by approximately 47% between late 2023 and late 2024, yet actual hires for those same roles fell by roughly 73% over the same period. "Ghost jobs" — postings companies maintain to project an image of growth — are widespread. This makes the entry point harder to find, but it does exist.

Concrete Strategies for Breaking Through

Cast a Wide Net Through Your (Real) Network

Approximately 26% of job placements come through referrals. Take stock of your actual network — classmates, professors, former internship contacts, family members — and identify anyone connected to companies that may be hiring. The goal is a warm introduction that reaches, or leads to, a decision-maker. One effective referral is worth more than a hundred cold applications submitted through job boards.

Look for Symmetric Risk

An entry-level engineer is inherently a high-risk hire. Startups carry a matching risk profile — potentially lower pay, no guarantee of longevity, and higher performance expectations. But that shared risk creates mutual incentive. The learning curve is steep, the exposure is broad, and the hands-on experience translates directly. For engineers whose long-term goal is a large organization, a startup stint is not a detour; it is a way to build the experience those larger organizations ultimately want to see. A first job is about validating capability and learning — not a lifetime commitment.

Create Experience Rather Than Wait for It

Employers demand experience but are unwilling to hire you to gain it. The way through is to create your own: deploy a project, contribute to open source, build a real application for a small business or a family member. Recruiters are skeptical of toy projects, but a deployed application that solves a genuine problem — paired with the ability to articulate your design decisions and the reasoning behind them — still sets candidates apart.

Develop Practical AI Engineering Skills, Not Just AI Tool Usage

Using Cursor or GitHub Copilot is now a baseline expectation. What actually differentiates candidates is a deeper level of capability. Most working engineers — including senior engineers — have never built a RAG (Retrieval-Augmented Generation) pipeline or designed a multi-agent system. Understanding how to chunk documents, generate embeddings, store and query them in a vector database, and wire everything into a production application puts a candidate well ahead of the market. This is precisely where demand is growing fastest.

AI and data science-related job postings grew 163% in 2025. Engineers who genuinely understand how these systems work — rather than simply knowing how to write prompts — are the scarcest talent in the market right now.

Stop Optimizing for What You Cannot Predict

Nobody foresaw the 2021 hiring boom, and nobody predicted this correction. Build durable skills. The market's demand for engineers who can reason clearly about systems is not going away. Your starting point is not your endpoint.

---

Further Reading

Meta and Microsoft Join the Layoff Wave — Is AI Really to Blame?

More major tech companies are facing significant workforce reductions: Meta has announced it will cut 10% of its workforce, approximately 8,000 employees, while Microsoft plans to offer voluntary buyouts to 7% of its U.S. staff. Many attribute these layoffs to AI, but is that the real cause? Two scholars from the University of Sydney have shared their analysis in The Conversation.

Proposed Ban on Chinese Robots: Another Step in the Push for U.S. Tech Sovereignty

The United States continues to push for restrictions on sensitive technology of Chinese origin across multiple industry sectors. Proposed legislation has now extended to ground-based robots, including humanoid robots, robot dogs, and tracked robotic platforms. This could benefit some U.S.-based robotics companies, though many still rely on Chinese-manufactured components. IEEE Spectrum technology policy editor Lucas Laursen writes: "The U.S. robotics industry is caught in a bind."

---

原文來源： 查看原文