• About
• Contact
• Resume
• Professional Experience
  • Open-Source Software developer
  • AR-HUD Software Engineer @ Volkswagen AG - CARIAD
• Projects
• Education
• Spoken Languages
• Activities and Interests

About

In a hurry? Consider reading the more compact, public CV version.

I am a software engineer specializing in the domain of computer graphics (i.e., writing or dealing with programs/tools/APIs for the purpose of rendering or/and simulation) and high performance computing (i.e., multi-threading, parallel programming, etc.). I have professional experience in the automotive industry in the area of Augmented Reality (AR) Head-up-Displays (HUD) (see AR-HUD Software Engineer @ Volkswagen AG for details).

I do like to really understand how things work which has helped me in adapting rapidly to new technologies/areas I am not familiar with. From time to time, I like to write what-I-think-are high-quality guides – you can check them here :-).

I have a keen interest for FOSS projects. I try to use them whenever possible to avoid vendor lock-ins, contribute back to the community, and promote better alternatives to shady-licensed software.

I am proficient in:

• C++: Graphics programming (Vulkan, OpenGL, SDL), network programming, and GUIs using: Qt, Dear ImGui, and WxWidgets
• Python: General purpose scripting, sockets network programming, fast prototyping, parsing, data analysis, and, God forbid, GUI programming using Tkinter.
• C#: Game Development using Unity Game Engine, back-end web development using ASP.Net, and other misc dotnet-based desktop applications
• TypeScript: Front-end rendering on the Web using: threejs, d3js, Deck.gl, etc.

I can somewhat program in:

• C: Writing projects to better understand low-level concepts
• Lua: Neovim configurations and professional use in automotive industry (proprietary Engine customization. See basemark’s Rocksolid AR)
• Bash: General purpose scripting tasks

Contact

In case you want to contact me then the best way to do so is via email walid.chtioui@ensi-uma.tn.

You can also contact me via LinkedIn (at: walid-chtioui) but I prefer email.

Resume

You can access my public resume in English here or in German here (In case of any difference between the two versions, the English version is the reference).

In case you require a signed resume with additional details (e.g., nationality, date of birth, etc.), contact me and I will provide it.

If you are looking for more details about my resume (i.e., projects that I worked on) then see Professional Experience section below, projects and guides I worked/am working on.

Professional Experience

Below is my professional experience in detail. In case you are in a hurry (or just looking for certain keywords), just take look at my resume.

Open-Source Software developer

Duration: 09/2025 - present

Mainly OSS contributions to GPU or computer-graphics-related major projects including KTX (the texture format), OpenImageIO (used extensively in animation and VFX studios all over the world to provided a common interface to reading and writing images), and Blender.

Some PRs I contributed (some of these require/required multiple extensive reviews to be accepted):

• KTX-Software #1167: Contributing BCn with rate distortion optimization (RDO) GPU compression to Khronos texture format (KTX).

• OpenImageIO #5185: Contributing to add KTX format reading/writing support to OpenImageIO (OIIO) (which in turn adds KTX support to Blender3D).

• KTX-Software #1223: DDS-to-KTX2 lossless converter (i.e., convert DDS texture files to modern KTX2 format without decoding then encoding again).

• KTX-Software #1205: Fix miscellaneous memory issues with libktx and KTX tools.

For other PRs/Issues that I have opened/worked on, see my KTX-Software PRs and KTX-Software issues.

AR-HUD Software Engineer @ Volkswagen AG - CARIAD

Duration: 04/2023 - 04/2025

I worked at CARIAD with an interdisciplinary AR-HUD team to, among other things, build tools, deploy and test proof of concepts, and write C++ samples.

Before I describe in-detail my work experience as an AR-HUD software engineer (well not entirely, but according to what I can - because of NDA and such), the concept of AR HUD has to be explained. What I am writing here is public knowledge information – VW, please do not sue me.

Heads-up-display (HUD), in the context of the automotive industry, projects information on the windscreen of the car in the direct field of view of the driver. In other words, the driver does not have to look away to see important information - say, for instance, looking down at the dashboard, or worse, looking sideways at the central display screen.

The above picture illustrates the two projection fields for the HUD:

• Status field (or near field) – mainly used for 2D status information (e.g., speed, current traffic sign, and general status messages). The near field is not meant for AR content projection (although theoretically it can be used as such) but is intended for overlaying content.
• AR field (or far field) – the larger field that is used for actual AR content (e.g., showing navigation arrows, showing current ACC, etc.)

This video showcases the AR-HUD feature/concept for the Audi Q6 e-tron that I contributed to at CARIAD.

• Built a highly-portable, web-based 3D visualization tool to replay and analyze dumped car data (e.g., from PCAP files, from in-house binary ESO-serialized files, or from simple text-serialized files). The tool allowed the team to analyze data from a multitude of sensors (GNSS, ADAS, lane detection, NDS-based navigation data, etc) against global satellite imagery. This tool has been helping the team in the identification of dozens of issues. Used technologies: Deck.gl, WebGL, and TypeScript.

  The visualization tool is essentially a web-based WebGl simulator/player that gets as input dumped data either from PCAP (packet captures using Tcpdump), ESO data (internal binary format provided by esolutions), or custom-provider-based formats. The input is parsed and visualized accordingly. Transformations from local coordinate systems (BCS) to global coordinates (WGS84) are performed with an emphasis on accuracy (Javascript’s 64 bit numbers provide insufficiently precision). The tool lays out data in local coordinate systems (e.g., BCS) against global satellite imagery. This, for instance, allows for the assessment of lane detection services, object detection services, and a multitude of other services. The tool is around 20 000+ lines of Python and Typescript code whose rendering part is mostly developed by me. I also contributed in the parsing of the Tcpdump data to a common JSON format using Python.

  Important to note that initially we tried to approach everything using only Python and we tried to use Folium (a Python wrapper around Leaflet) but discovered quickly that it was very limited. I then proceeded to write my own Python wrapper which initially proved to be helpful but then also suffered from limitations because of the underlying 2D, non-GPU accelerated Leaflet map renderer.

• Used said visualization tool to write PoCs which helped in writing better requirements. Used technologies: DOORS.

  To save costs and decrease development times, some proof of concepts (PoCs) are developed using said visualization tool and are checked for validity using pre-captured data.

  This approach resulted in us figuring out that certain features/functionalities had some issues when tested on real test drive data. Consequently, requirements were adjusted before final submission to the provider(s) which avoided potential future issues and significant delays.

• Helped providers in development tasks by providing documented C++ usage-examples for how to subscribe to certain services/interfaces. Used technologies: C++, CMake, SomeIP, OpenGL, OpenGL ES, IP Sockets.

  Particularly, I contributed in the initial setup of HUD devices by providing C++ sample code on how to render a triangle (i.e., the usual Hello World of computer graphics but for AR-HUDs rather than conventional 2D screens) on the provided HUD device. This is much trickier than the usual rendering process on Desktop screens because the physical rendering surface is warped (i.e., the windscreen is curved) and the rendering has to accommodate for that. There are additionally a lot of Gotchas. E.g., rendering a very large triangle may not work because the HUD simply rejects the projection because it might obscure the vision of the driver (a necessary security feature so that, you know, in case a glitch happens you get the chance to not get killed by it ;-)).

  This task required some roll-your-sleeves-and-get-your-hands-dirty type of debugging where documentation was sparse (or worse – non existent).

• Contributed to identifying and solving issues related with certain AR-HUD functionalities. E.g., Audi’s global drone positioning improvements (see the blue arrows in the picture below):

  

  Positioning a global element (in this case the turn icon) in the body coordinate system (BCS) of the car is quite challenging and heavily dependent on the accuracy of global positioning sensors and their data fusion (e.g., GPS), the accuracy of navigation data (e.g., NDS), local orientation sensors (e.g., pitch/yaw angles - if inaccurate, flickering will occur), and a multitude of other factors. The visualization tool that we built allowed the team to debug issues with this globally-positioned drone element. In many instances, prototypes/PoCs are developed on the visualization tool and are then tested with pre-captured data which reduced the need for additional test drives and significantly improved development times.

• Contributed to writing requirements about navigation-related, NDS-based functionalities.

• Wrote Python scripts to parse and visualize a multitude of data dumps. Used technologies: Python, Tkinter, Plotly.

  Important to note here that this made me realize how absolutely inefficient Plotly is for anything out of the usual, generic visualizations. Also, GUI programming in Python is something to avoid like the Plague.

• Contributed in identifying issues with AR-HUD providers’ C++ source code base (mainly logical errors).

  Having to do significant jumps to debug the most trivial issues is surely NOT an indication of a high-quality C++ codebase. Also, using design patterns for the sake of just having design patterns is extremely frustrating.

• Worked within a highly interdisciplinary team involving software engineers, product managers, and test engineers.

  This in turn required that the tools we build are easily usable (therefore portable) by laymen with a wide range of expertise (from expert software engineers to other colleagues that are not necessarily IT inclined).

Projects

See projects section for a list of projects I worked/am working on.

Note: there is no LLM generated code in these projects. Most of them are done for learning purposes and using LLMs defeats that.

Education

In April 2025 I obtained a double Master’s degree in computer science at the University of Passau (Germany) and the National School of Computer Science (ENSI) (Tunisia).

My double Masters graduation thesis is titled:

Direct Volume Rendering of Large Volumetric Datasets in Immersive Environments

Thesis’ source code is fully accessible and documented here.

Relevant coursework:

• Randomized Algorithms
• Mathematical Foundations of Machine Learning
• Implementation of Cryptographic Algorithms
• Principles of AI Engineering
• Introduction to Deep Learning
• Wireless Security
• Responsible Machine Learning (Interpretable Machine Learning)
• Advanced Topics in Data Science
• Immersive Analytics
• Project in Visual Computing
• Operating Systems
• Operations Theory
• Statistics
• Computer Architecture (MIPs, x86 Assembly)
• C++\C Programming Courses
• Complexity Theory
• Probability Theory

Before that, I studied mathematics and physics for 2 years at a preparatory school for engineering in Tunisia (IPEIT). In which I passed a nation-wide examination for entering engineering schools. I achieved an average-ish rank of 145 over 2200+ participants.

Relevant coursework:

• Algebra (with extensive focus on linear algebra)
• Analysis (with extensive focus on topology)
• Physics (EM waves, Maxwell equations, introduction to quantum physics, etc.)
• Control Theory
• Python (2.7 because God forbid a curriculum updates a technology it teaches :-/)
• SQL databases

Spoken Languages

• English – C1 (fluent)
• French – C1 (fluent)
• German – B1 (beginner-intermediate)
• Arabic – Native

Activities and Interests

In case you are a recruiter reading this and you happen to play Volleyball and/or Football, consider adding me to your team :-D.

Other than sports, whenever I have free time, I like to read technical books or improve my German language skills.