These famous tech careers all started with tiny side projects
Side projects usually seem like a hobby, but they can be the most powerful career launching pad available. While traditional metrics like college degrees and years of professional experience still carry weight, a well-executed side project acts as a direct, unfiltered view of your technical abilities, problem-solving approach, and passion for creating.
These projects didn’t just fill space on their creators’ CVs. Their work has created new standards and made them indispensable thought leaders in their fields.
Kim Swift and portal
The video game Portal was born from a student side project called Narbacular Drop. It was created by DigiPen Institute of Technology students, including Kim Swift, for their senior project. The main idea was a 3D platforming mechanic in which players used portals to move around the environment. The team showed the game to recruiters at DigiPen’s annual career day, where it was viewed by Robin Walker of Valve, the company that makes the Steam Machine. This led to an invitation to present at Valve’s offices.
They didn’t know that Gabe Newell, the co-founder of Valve, would show up, but they introduced him to him anyway. About 15 minutes into the demonstration, Newell stood up and called for everything to stop. He offered the entire team of seven people to work on site. They were hired to rebuild the game under the name Portal.
Ari Weinstein and the Workflow app
Ari Weinstein built Workflow with a team at the MHacks Winter hackathon. Weinstein, a Thiel Fellow who dropped out of MIT, designed Workflow to automate complex tasks on iOS devices, filling a huge gap in mobile automation. The project quickly gained critical acclaim, winning an award, and led to Apple acquiring the Workflow app in March 2017. This led to Weinstein and his team being hired.
This “purchase-rental” allowed Apple to integrate the technology directly into the operating system. Workflow has been transformed into an iPhone shortcuts app. Weinstein was head of software engineering at Apple for six years, overseeing the development of shortcuts and other proactive intelligence features until his departure in 2023.
Kenneth Reitz and Python queries
Kenneth Reitz, photographer and developer, started the Requests library as a side project. Frustrated with the complexity of HTTP interactions in Python, he struggled with the standard urllib2 module when creating a Convore API wrapper. This module required unnecessary hoops and abstractions. He built the first prototype of Requests in about two hours.
The developer community immediately loved the project and it became one of the most downloaded Python packages. This success was a catalyst for Reitz’s career. Heroku, a major cloud platform, hired him as a Python architect for his open source contributions. Having the freedom to choose his own job title, he chose “Python Overlord”. In this role, he was in charge of the entire Python experience on the platform, continuing to build supporting tools.
Dan Abramov and Redux
In 2015, Dan Abramov wanted to give a talk at React Europe called Live React: Hot Reloading with Time Travel. His problem was that he didn’t know how to implement the time travel function. To solve it, he looked for a state management model that would allow actions to be replayed and code exchanged. Drawing on ideas from Elm architecture and existing Flux libraries, he created Redux as a proof of concept for his conference demo. The presentation was a huge success, demonstrating a new state management model that quickly outperformed existing Flux configurations.
After his speech, Abramov met React team member Jing Chen in the audience. When she asked if he was interested in working at Facebook, Chen suggested the London office and immediately arranged an in-person interview at the conference hotel in Paris. He joined Facebook’s React Core team and his side project became the industry standard library for state management in React applications.
François Chollet and Keras
François Chollet created the Keras deep learning library in early 2015 as a side project for his research into recurrent neural networks and natural language processing. Chollet prioritized developer experience and low cognitive load, designing Keras as a high-level, intuitive interface capable of running on different backends.
Shortly after Keras’ release, Chollet was hired by Google to research computer vision. Keras remained his personal project until Google publicly released TensorFlow. Chollet refactored Keras to support TensorFlow as a backend engine, which led TensorFlow team lead Rajat Monga to find it. In 2016, Monga offered Chollet a position on the TensorFlow team to officially integrate Keras into the platform. Keras is now the official top-level API for TensorFlow, making it the frontend for one of the most dominant machine learning platforms in the world.
Chris Lattner and LLVM
Chris Lattner started the Low Level Virtual Machine (LLVM) project as part of his master’s thesis and doctoral research at the University of Illinois at Urbana-Champaign in the early 2000s. It was released as open source in 2003, providing a new modular alternative to the GNU Monolithic Compiler Collection (GCC), which rejected LLVM as a backend.
Lattner’s work on this complex infrastructure caught the attention of Apple, which was struggling with the limitations of its existing compilation tools. His open source efforts built relationships with Apple engineers who saw the revolutionary potential in their developer ecosystem. Apple hired Lattner after he completed his Ph.D. in 2005 to develop LLVM full-time. Apple integrated LLVM into its core development stack, and Lattner created the Clang compiler and Swift programming language.
Sébastien McKenzie and Babel
Sebastian McKenzie started the project that became Babel while he was a high school student in Australia. He wanted to understand how parsers and compilers worked. It worked as a transpiler, allowing developers to write new code using upcoming standards and convert it to the widely supported JavaScript ES5 format. The project took off immediately because it allowed engineers to use modern syntax without waiting for slow browser vendors to implement support.
The project’s rapid adoption and quality gave him a powerful portfolio, attracting the attention of major tech companies despite his lack of a college degree or professional experience. After working at Thinkmill and Cloudflare, McKenzie was hired by Meta in 2015. The company recruited him to maintain Babel and manage their internal JavaScript infrastructure, effectively paying him to develop its open source tool.
The general principle is that these projects were not unified by luck, but by a relentless commitment to quality design, regardless of whether it would sell. Sometimes all it takes is for a project to be useful for a company to take notice.
So if you have a side project, it’s probably a good idea to treat it like it really matters. Stop waiting for planning permission and instead choose a problem that really frustrates you. Solve it with excellence, put it out into the world, and your code could become the ultimate career springboard for you too.
