Arman Hezarkhani

Arman Hezarkhani is a co-founder and managing partner of Tenex, alongside Alex Lieberman. His work combines AI-assisted software delivery with an output-linked compensation model: clients pay for delivered story points, and engineers earn variable compensation for completed, accepted work. In his AI Engineer Code 2025 talk, he argued that engineers should receive a direct reward for using AI tools well while maintaining high standards of code quality.[3][2]

Hezarkhani frames AI adoption as an incentive problem. In his talk, he recalled working at the 33 Irving WeWork in New York and seeing a neighboring engineer type directly into Visual Studio Code, without a chat window. He contrasted that with his own search for AI tools that could improve speed and quality: as a founder, he benefited from incremental efficiency gains. His explanation of the neighbor’s salary, bonus and equity was speculative, but the encounter supplied a concrete illustration of his question: who benefits when an engineer finishes faster?[3]

Tenex’s model retains a flat base payment, then adds a fee per completed story point when a ticket is accepted, with a quarterly true-up. Strategists distill product requirements with clients; engineers turn those requirements into architecture design documents containing tickets and story-point estimates before implementing the code. Hezarkhani says architecture design takes most of the company’s engineering time. The variable reward therefore depends on accepted tickets rather than hours spent writing code.[3]

Hezarkhani acknowledges three risks: inflated story points, rushed work that lowers quality, and unhealthy competition among engineers. Tenex counterbalances output incentives with strategists paid according to net revenue retention, reviewed scoping, multiple internal quality-assurance rounds and client approval for every ticket. For collaboration, he emphasizes hiring judgment: “We make hiring incredibly difficult for ourselves so that everything else is easy.” He also credits Lieberman with the metaphor that AI magnifies an engineer’s existing strengths—or sloppiness.[3]

In a November 26, 2025 post, Hezarkhani summarized his priorities as “Output > hours” and “Impact > tenure.” He argues that conventional compensation can fail to reward experimentation and effective AI use, but stops short of prescribing Tenex’s arrangement for every team: “I’m not saying everyone should adopt our model.”[4]

Hezarkhani illustrates the delivery model with Tenex’s work for a billboard company whose self-service customers uploaded images for advertisements. Images required internal moderation and approval from billboard owners; he identified the internal review as both an operating cost and a delay before advertisements could run. He reports that Tenex scoped, architected and built an AI moderation system in two weeks, reaching 96% accuracy compared with a human moderator. That reported comparison came without a sample size or evaluation protocol.[3]

For another client, Hezarkhani describes expanding the capabilities of retail-camera devices that could initially run one heat-mapping model. Tenex explored additional chip capacity and model quantization to support parallel execution, then, he reports, built five models that could run concurrently, including heat mapping, queue detection and theft detection. He presents this as team work following the same requirements-to-architecture-to-implementation process.[3]

Before Tenex, Hezarkhani worked in education technology, consulting and personal finance. In his personal biography, he describes studying at Carnegie Mellon University and taking his first education-technology startup through AlphaLab, followed by two internships at Google for Education. He then founded Ahez Consulting, helping Fortune 500 companies use education to market high-technology products. In his 2025 talk, he also recalls teaching in Carnegie Mellon’s School of Computer Science and helping Google scale its AI, cloud and mobile practices internationally before founding venture-backed startups.[6][3]

Hezarkhani founded Parthean, whose stated mission is to make financial information universally accessible. Its team and investor page lists him as founder and CEO and names Lieberman among its investors. Their relationship thus predates their collaboration at Tenex, connecting Hezarkhani’s earlier startup work with his later AI engineering venture.[5][6][3]