Tereza Tížková

Tereza Tížková works on growth at Factory, the AI-native software development platform whose autonomous coding agents, Droids, work across the development lifecycle. Previously E2B’s first hire, she helped developers understand and use sandboxed computers for AI agents. Her work combines developer education, product positioning, and community building with an interest in what agents need to accomplish useful work beyond generating code.

From mathematics to developer education

Tížková studied mathematical structures at Charles University in Prague and worked in management consulting at McKinsey before entering AI infrastructure. During the pandemic, she founded a local IAESTE branch to organize international technical internships for students. She joined E2B as its first hire in June 2023.

At E2B, her growth responsibilities crossed developer content, partnerships, analytics, launches, brand, and community. Explaining the product also meant learning Python and TypeScript to create examples herself: developers needed to see what they could build with a sandbox, not simply hear that the infrastructure existed.

She built and maintained Awesome AI Agents, an open-source directory of autonomous-agent projects. The directory gave developers a way to explore the emerging ecosystem through concrete projects and repositories, while introducing E2B through a useful resource.

She also co-created and co-taught Building Coding Agents with Tool Execution with E2B machine-learning engineer Francesco Zuppichini. The course teaches agents to generate and safely execute code in cloud sandboxes, extending what they can do beyond a predefined menu of function calls. It connects an infrastructure product to a practical capability: giving an agent somewhere to run the code it writes.

Her subsequent growth role at Factory connects that execution infrastructure to software-development workflows. E2B provides environments in which agents can act; Factory coordinates agents that build and validate software. Her writing explores the requirements connecting those two layers, including tool access, persistent state, and checks that establish whether the work succeeds.

Making agent systems understandable and useful

• Growth through useful explanations: Tížková treats growth as finding and removing the obstacle currently holding a company back. In her account of startup growth work, the response might be a coded example, a partner integration, a launch, or clearer positioning. She pairs creative work with analytics: tracing where users arrive from, observing how they navigate a website, and comparing results with the resources an activity consumes. Producing some benefit is insufficient reason to keep investing in an initiative.

• A computer is part of the agent: Her writing about Manus and E2B makes the execution environment central to understanding an agent. Planning and generating code are only part of the workflow; an agent also needs to browse, install dependencies, manipulate files, and preserve progress between actions. Her computer-use explainer separates the model, the surrounding agent software, and the machine it operates. That distinction explains why a capable model alone does not produce a dependable autonomous system.

• Verification through actual use: Tížková emphasizes testing an application’s behavior, including interacting with its interface. Her computer-use writing explains how structured browser and accessibility information can guide agents, while different applications require different control methods. Persistence matters too: an agent handling a pull request must be able to wait for checks, respond to feedback, and resume its work. In her account of software factories, Factory’s validators assess code they did not write against completion criteria established before implementation. A user-testing validator clicks through the application on a virtual computer, checking whether the result works in practice.

• Software abundance changes what customers value: In her argument about software becoming marketing, Tížková predicts that easier software creation will weaken the position of interchangeable tools and make trust, distribution, domain expertise, and delivered outcomes more consequential. Her accounting example captures the distinction: customers may ultimately pay for their books to be closed, rather than for another application that helps them do it. This is a forecast about how software markets could develop, with implications for engineers and growth teams.

Autonomy depends on the surrounding organization

Tížková’s software-factory argument extends these ideas across the development lifecycle: collecting user feedback and logs, prioritizing work, executing it, validating results, and improving the process. She places much of the difficulty in the coordination and verification around code generation. An agent running for a long time does not by itself establish reliability; poorly chosen completion criteria can reward passing tests without accomplishing the intended task.

She describes Factory’s mechanisms for managing that work without claiming them as individual inventions. Automatic model routing classifies task difficulty and selects the cheapest model predicted to meet the required capability threshold. Factory Missions hand work between sequential workers with fresh context, then use independent validators to assess the result. Deferred context gives agents short tool descriptions first and loads full details when needed, limiting the amount of information competing for their attention.

Her organizational advice is equally concrete. Reproducible environments, useful tests, and clear documentation prepare a codebase for agents; without that preparation, she warns that automation can compound existing disorder. Plugins package reusable skills and team knowledge so people need not repeatedly explain how work should be done. Her preferred direction is for humans to decide what software to build while agents take on more implementation and routine coordination—a position she presents as an ambition for software factories, rather than a settled outcome.

Outside company roles, Tížková’s projects include teaching mathematics at Discover Summer Academy and organizing The Growth Lab, a community for startup growth practitioners. Her writing also spans topology and Gödel’s incompleteness theorems. These activities continue her interest in making difficult ideas usable, whether someone is learning abstract mathematics, choosing an agent tool, or bringing a technical product to market.