lukas rambold ✌️

Hi, I’m Lukas. Currently, I’m a master student of IT System Engineering at HPI in Berlin, focusing on HCI & Personal Fabrication.

Previously, I studied at LMU & CDTM and co-founded the government data-analytics startup Polyteia.

Mail · Google Scholar · LinkedIn


supervisor: Robert Kovacs

System for Designing and Fabricating Dynamic Structures

In this work, we are building a system that allows novice users to create stable structures that are subject to dynamic forces and acceleration. I am responsible for the underlying simulation that is relying on differential equation solvers. For that, I looked deeply into continuous-time Modeling & Simulation systems like Modelica, current developments in differential equation solvers e.g. in Julia and non-linear optimization approaches.

to appear at UIST 2021

supervisor: Thijs Roumen

Automatic Assembly of 2D SVG Plates to 3D Models

Laser cut models are commonly shared as 2D vector information (usually SVGs). This makes it hard to alter the underlying implicit 3D model. In this work we are introducing a tool that allows users to reconstruct the 3D model from vector plate data and therefore, thus making parametric changes possible. I was focusing on replacing the original joint geometry and generate new joints to match different kerfs, material thickness and geometry changes as well as visualizing the assembly tree search algorithm to enable parameter tuning.

project results were published at CHI 2021 & UIST 2021

supevisors: Christian Tietz, Eric Klieme

Under Pressure: Behavioral Authentication on Door Handles

Behavioral Authentication is wildly used in websites and smartphones. But how does it translate to other domains of all-day life? In this work, we equipped a door handle with resistive and capacitive touch sensors and found out in a study that we can authenticate a person pressing the handle with 88% accuracy by classifying the sensor data. I contributed the initial concept, created the Arduino-based hardware protoype, conducted the user study and implemented the classification code in Python.

published at SECRYPT 2020 (best poster award)

startup as technical Co-Founder

Polyteia: Enabling Data-Driven Decision Making in Municipalities

We founded the startup Polyteia to help mayors and city councils to have their municipalities data in one place to analyze and compare. As the technical Co-Founder, I hired and manged the engineering team, lead the development of the custom data visualization components in D3js and custom data integration solution using Python, React & Electron and led publicly funded R&D projects.

finalist at TechCrunch Distrupt 2018 and featured in FAZ, Zeit & others

supervisor: Patrick Baudisch

Assembly Instruction Generation for Laser Cut Models

Since laser cutters are quite fast, assembly is becoming the new bottleneck in the fabrication process. To address this, we built an instruction booklet generator for arbitrary plate models. I was responsible for rendering the images and layouting the final PDFs on-the-fly using TypeScript and Three.js.

now part of Kyub

supervisor: Christian Mai

TransparentHMD: Overcoming the Communication Barrier of HMDs

Many people report that when there is one person in the room who uses a VR headset, it tends to become a disconnecting experience for the others. In my bachelor thesis, I worked on improving collaboration between in-VR users and bystanders by displaying the in-VR user’s face on top of the HMD. Additionally, the system tracks the eyes of its observer to correct for the parallax error of the 2D display on the outside of the HMD. As part of my Bachelor thesis, I implemented the fully functional prototype in Unity and Android and conducted a user study to explore the effects the system has on collaborative tasks.

published at MUM 2017; concept was cited and adopted by Facebook Research