# Eitan Feldman > I build things that run. Products, software, robots and companies. Based in Buenos Aires. 17 years old. Still in high school. ## Nine layers deep "Full-stack" usually means two layers and gets said like it is a lot. A real system runs from the sensor that measures the world to the decision about who pays. No single project below covers all of it. Together they do. - **08 Business**: CEO, team of 4, first venue live (Restoman, Águila) - **07 Product**: specs, design system, RBAC model (Restoman, Gemm) - **06 Interface**: Next 16 · Tailwind 4 · 4 apps (Restoman, Gemm, Águila) - **05 API**: orders per context, permissions (Restoman, Águila) - **04 Data**: schema, migrations, cache (Restoman, Águila) - **03 Infra**: haloy · GitHub Actions · CI (Restoman, Gemm) - **02 Systems**: Swift, LaunchDaemon, code signing (Awake) - **01 Perception**: SLAM · FAST-LIVO2 · ROS2 (Gemm) - **00 Physical**: Unitree G1 · LiDAR · refinery round (Gemm) ## Selected work Different problems. Same mindset. From idea to the real world. ### [Restoman](https://eitanf.com/work/restoman.md) **Restaurant infrastructure, built differently.** Restoman is a B2B platform that runs a restaurant's floor — orders, tables, kitchen and checkout — built by a team of four that I lead as CEO. It is live in its first venue, and it started as a school project that only worked for one restaurant. Role: CEO. Product, design, and most of the code. · Period: 2025 — now · Status: Active ### [Gemm](https://eitanf.com/work/gemm.md) **Autonomous industrial robotics.** Gemm is building a humanoid that can walk a refinery round a person currently walks every three hours: reach a furnace, read its gauge, and turn a valve based on what it reads. I own the data bridge, the real-time 3D and 2D mapping, and the dashboard. Role: Backend and AI research. Data bridge, mapping and dashboard. · Period: 2026 — now · Status: Active ### [Águila](https://eitanf.com/work/aguila.md) **We built it. We launched it. We learned. We shut it down.** Águila was a platform where students found restaurant promotions near school. It reached ten paying restaurants in a few months. When the team decided to close it, we refunded every one of them — none of them asked us to. It is the shortest project on this site and the one I learned most from. Role: Code. Co-founder with one person on sales and one on design. · Period: 2025 · Status: Closed ## Right now - Building Gemm — perception for Unitree robots (Mar 2026) - Growing Restoman — first venue live (2025) - Finishing secondary school (until Dec) - Writing about what I learn (ongoing) ## The lab Small things, experiments and open source. - [This site](https://eitanf.com/lab/this-site.md): Built in the open: Next 16, no dark mode, monochrome, and a rock that is a point cloud rendered in about 6 KB of hand-written WebGL instead of 150 KB of three.js. - [Awake](https://eitanf.com/lab/awake.md): Awake is a macOS app that stops a Mac from sleeping, from the menu bar or the terminal. One binary, four modes, published as an open source release. - [Gemm Blocks](https://eitanf.com/lab/gemm-blocks.md): Visual block programming for controlling Unitree robots without writing code. It is what we worked on until we landed on the final Gemm idea and the G1 arrived. Shown at a hackathon, which turned into an invitation to teach a class about it at ITBA. - [Sana](https://eitanf.com/lab/sana.md): Sana is an AI triage concept for places where the nearest specialist is six hours away. Designed, prototyped and pitched in two hours at a hackathon at Universidad Torcuato Di Tella. - [Horario ICS](https://eitanf.com/lab/horario-ics.md): Photograph your school timetable, get a subscribable calendar. Gemini does the OCR and the structured extraction; the app turns the result into an .ics feed your phone can follow. ## Notes Decisions, lessons, and what I take away from others. - [The adapter was the bottleneck, not the SLAM](https://eitanf.com/notes/the-adapter-was-the-bottleneck.md) (8 September 2026): We spent weeks optimising a SLAM pipeline whose real limit was the layer translating between two message formats. The lesson is about where you point a profiler, not about SLAM. - [We took out the auth plugin and wrote our own RBAC](https://eitanf.com/notes/we-wrote-our-own-rbac.md) (22 August 2026): Better Auth's organizations plugin models a flat company. A restaurant group is a hierarchy, and permissions have to inherit down it while never escalating up. Bending the plugin would have cost more than replacing it. - [Why Awake's helper is a LaunchDaemon and not a sudoers rule](https://eitanf.com/notes/launchdaemon-not-sudoers.md) (27 July 2026): The whole point of the app is asking for a password once instead of every time. Only one of the two options survives a reboot, and that decides it. - [Seven green deploys that never shipped](https://eitanf.com/notes/seven-green-deploys.md) (13 July 2026): Restoman's API served the previous night's build for a day while seven deploys reported success. A stuck migration and a deploy command that exited 0 on failure hid each other. - [We went to learn someone else's tool and left with a class to teach at ITBA](https://eitanf.com/notes/hackathon-itba-class.md) (3 June 2026): A Google hackathon at ORT was about making videos with Google Vids. We used the room to show Gemm Blocks instead, which led to an invitation to teach it at ITBA. - [Two hours to design triage—and most of the work was defining limits](https://eitanf.com/notes/two-hours-to-design-triage.md) (11 May 2026): A hackathon at Di Tella asked how AI could help where the nearest specialist is six hours away. Our answer was defined less by its features than by the lines it would not cross. - [What I took away from Daniel Rabinovich's talk at Mercado Libre](https://eitanf.com/notes/mercado-libre-visit.md) (15 April 2026): With ORT, we visited Mercado Libre's offices. Four ideas on the impact of artificial intelligence and the future of talent, and a call to get in the game. ## About I build companies, software and robots while finishing secondary school in Buenos Aires. > Take it seriously before anyone else does. Even when it starts as an experiment. If it's only for learning, it stays an experiment. [More about me](https://eitanf.com/about.md) ## Contact - Email: me@eitanf.com - GitHub: https://github.com/eitan-que - LinkedIn: https://www.linkedin.com/in/eitan-feldman/ - Resume: https://eitanf.com/resume.md --- Page: https://eitanf.com/ · Español: https://eitanf.com/es/index.md · llms.txt: https://eitanf.com/llms.txt