## Notes from 09 April 2026 [[2026-04-08|← Previous note]] ┃ [[2026-04-10|Next note →]] # [[US Tech Force]] and the Talent Allocation Problem The **[US Tech Force](https://techforce.gov/)** is a federal hiring initiative launched by the [[US Office of Personnel Management (OPM)]] in December 2025 under the Trump administration. The program aims to recruit approximately 1,000 technologists annually—primarily early-career software engineers, AI specialists, and data scientists—into federal agencies for one- to two-year fellowship terms. Fellows will be supervised by managers recruited from the private sector, with teams reporting directly to agency heads. OPM Director [[Scott Kupor]] has framed the initiative as a way to create a "fluid career track" between government and the private sector, noting that over [35,000 individuals expressed initial interest](https://federalnewsnetwork.com/hiring-retention/2026/01/opm-extends-tech-force-application-deadline-citing-tremendous-interest/). The program is run [with the support](https://www.linkedin.com/feed/update/urn:li:activity:7422277036867588096/) of [[Amanda Scales]], who brings extensive experience in talent management from Uber (8 years) and Musk’s companies (SpaceX, The Boring Company, and xAI). She has already secured partnerships with major tech firms including AWS, Meta, Microsoft, Nvidia, and Oracle. Scheduled to begin its first cohort by September 2026, the initiative follows the Trump administration's dismantling of previous federal digital service teams, such as the [[US Digital Service (USDS)]], [[18F]] at [[General Services Administration (GSA)|GSA]], and parts of the [[DHS AI Corps]]. --- **[[Abigail Haddad]]**, a data scientist and ML engineer with a Public Policy PhD from [[RAND Corporation|RAND]], was formerly part of the DHS AI Corps—at the time, the largest AI team in the US federal government. Now in the private sector, she published an opinion piece in FedScoop on February 20, 2026, titled _"[How to Make Tech Force Work](https://fedscoop.com/tech-force-strategy-blueprint-for-opm/)"_. The article serves as a practitioner’s blueprint for success, highlighting how the program might fail if lessons from previous efforts are ignored. This is one of the best pieces I have read on government technology talent programs. It resonates deeply with my experience at [[Motriz|Vetor Brasil]], where the "talent allocation" problem was our daily bread. What makes the article exceptional is that Haddad avoids writing a mere critique; instead, she provides an operating manual. Notably, while her own team at the DHS AI Corps was dismantled by the very administration now launching Tech Force, she refuses to let that irony drive her argument. The result is something far more useful than a grievance piece. ### The real bottleneck is not talent... but the organization of work The article’s central argument is vital: the U.S. federal government does not primarily suffer from a shortage of technical talent, but rather a lack of the organizational infrastructure needed to absorb and deploy that talent effectively. Agencies often lack a backlog of well-scoped projects with clear roadmaps, defined deliverables, and executive sponsors. Instead, the default state is one of ambiguity—vague mandates, unclear ownership, and no success criteria. This is the insight that anyone designing a government talent program must internalize. Recruiting is the visible, marketable side of the work, but the invisible, unglamorous task of structuring demand is what determines results. Without a prepared environment, programs risk producing what Haddad calls "innovation theater": fancy presentations and unread reports that result in no actual change for citizens. The problem is compounded by short tenures. A fellow on a two-year term who spends six months navigating bureaucracy has already lost a quarter of their time. Haddad argues that project scoping should be a full-time function at OPM _now_, months before the first cohort arrives. This is the precondition for any engineering work to matter later. I witnessed this repeatedly at Vetor and elsewhere: a placement is only as good as the project. When we placed a trainee with a defined problem and a committed sponsor, the results were excellent. When the mandate was a vague call to "bring innovation," trainees spent months in a bureaucratic fog, leaving frustrated and without tangible outputs. The lesson is universal: **talent without a prepared receiving environment produces frustration, not results.** As a general career rule, your job is often defined by the people you work with—specifically your boss. Choose the people first, then the challenge. ### The political economy of incoming talent Haddad is equally sharp on power dynamics. Tech Force fellows will report directly to agency heads, bypassing existing chains of command. This is inherently threatening. In many agencies, the technical workforce consists of contractors incentivized to control knowledge to ensure contract renewals. Meanwhile, career civil servants, demoralized by the layoffs of 2025, may view Tech Force as a surveillance mechanism or a prelude to further cuts. Individual goodwill cannot overcome a structural misalignment of incentives. Organizational arrangements must be designed so that cooperation is in everyone's interest. We saw this at Vetor Brasil as well; every external placement disrupts an existing equilibrium. Whether incumbents welcome a newcomer or protect their turf depends less on virtue and more on whether the organizational design gives them a reason to collaborate. ### Program design choices and their implications Two specific features of Tech Force warrant attention: 1. **Engineering Focus:** Unlike USDS and 18F, which included product managers and designers, Tech Force is engineering-heavy. This places the burden of scoping and stakeholder alignment elsewhere. If no one fills that gap, engineers risk "having opinions about other people's code" rather than writing their own. 2. **Private-Sector Supervision:** Bringing in managers from the private sector is a bold choice. However, Haddad warns that if these managers believe "good engineers fix things" by default, they will be blindsided. Government engineers often fail to deliver not due to lack of skill, but due to lack of access and restrictive decision-making flows. These managers must understand federal procurement and "authority-to-operate" processes _before_ they start. Haddad recommends they consult with alumni of previous digital service efforts. While this knowledge isn't always perfectly documented (though some efforts exist, [[2025-06-13|like the USDS oral history]]), there is a wealth of tacit knowledge among those who have run similar initiatives. ### What this means beyond the US The dynamics Haddad describes are not uniquely American; I see the same patterns globally. While governments certainly need more tech talent, the primary bottleneck remains the capacity of public organizations to define meaningful work and provide adequate tools. The assumption that tech talent is a magic wand is naive. The hard work lies in organizational preparation - and that work is never as politically rewarding as a high-profile launch event with corporate partners.