i build systems, untangle messy problems, and write code.
scroll
no rigid title. some days that's leading a team scaling a startup zero-to-one. others it's wiring agentic ai, building spatial interfaces in unreal, or tearing apart hardware just to understand it. the more impossible, the more interesting.
Eliáš did an amazing job on our project for Colin Stetson's new album, The Love It Took to Leave You. He took extra care at each step to ensure proper completion — which built trust and faith as a new client. The output was amazing: the animation of the album cover for Apple Motion was clean, simple and impactful, and the AR filter that works with the vinyl cover is a feast for the eyes. The longer we looked at it, the more we liked it. The details coming into frame, subtly and tactfully, show Eli's experience and good eye. We couldn't be happier with the final result and would happily work with him again.
Lisa WellsRedbird Musicredbirdmusic.com On set he was the calm one. Chrome ball, colour chart, HDRI — documented so cleanly that post never had to call us back. That kind of quiet thoroughness is rarer than it should be.
PlaceholderWilma Filmplaceholder He took a vague brief and a small budget and came back with a working MVP and a team that believed in it. We closed the round on the strength of what he'd built.
PlaceholderFlipasplaceholder We booked him to talk about spatial computing and he rewired how half the room thought about it. No slides full of buzzwords — just the work, and why it matters.
PlaceholderCzechVRFestplaceholder
- foundersombra os * proactive cognitive partner
- cto * 0 → 1flipas * lucid
- spatial computingrevizor * unreal * osint
- vfx supervisorwilma film * hagen
- speakernike berlin * czechvrfest
- vr / xrten years deep
01
pragmatism over purity
the best code ships, scales, and the next person can read it. right tool for the job — typescript, c++, python, doesn't matter.
02
the black box protocol
i'm drawn to systems i don't understand yet. break them to parts, find the edges, rebuild. fastest way to learn any domain.
03
agentic execution
point me at a fuzzy, structurally hard problem and walk away. high-ambiguity is where i'm most useful.
013d * vfx * xr
- blender
- unreal engine
- houdini
- three.js * r3f
- webgl * webgpu
- unity
- substance
- webxr
- virtual production
- glsl
02engineering
- typescript
- react * next.js
- node.js
- python
- postgres
- docker
- c++ * c#
- graphql
- aws * gcp
- systems design
03ai
- llms * prompting
- agentic systems
- rag
- computer vision
- pytorch
- neural nets
- nlp
- yolo
- langchain
- embeddings