About
Building digital humans since 2022.
A team of 11 with roots in medical computer vision. We make AI avatars usable in the real world — in lobbies, on kiosks, and inside closed government networks.
Disciplines
One team, research through production.
The same people build the offline avatar-creation pipeline and the real-time runtime. No hand-off between vendors halfway through.
Data scientists
Computer vision and medical imaging backgrounds. Peer-reviewed publications at MICCAI and IEEE, split between academic and industrial research tracks.
Security engineer
On-prem architecture, certificate authority design, mTLS device certificate issuance, audit pipelines.
Real-time engineers
Unreal Engine 5, real-time rendering, WebRTC streaming, custom shader work.
Animator
Facial rigging, blendshape design, motion authoring for character avatars.
3D artist
Modelling, rigging, texturing — owns the asset pipeline end to end.
Capabilities
What we take on.
Beyond our own platform, we take on projects as an outsource engagement, a dedicated pod, or turnkey delivery.
AI avatars and digital humans — research through production
Real-time AI and CV pipelines — low-latency, GPU-optimised
Computer vision across verticals — medical, retail, industrial QA, microscopy
On-prem AI deployment for regulated industries — isolated stacks, audit-grade
WebRTC and streaming media at production scale
UE5 and 3D content pipelines for interactive products
MLOps and infrastructure-as-code for AI workloads
Track record
Shipped inside closed networks.
We have twice run a government on-prem deployment end to end — from infrastructure groundwork to sign-off by the end user.
Government agency, Gulf region
Isolated contour: internal certificate authority, mTLS, encrypted model weights, national identity system integration, and self-service kiosks at customer-facing points. Arabic language support with speech recognition tuned for the regional dialect.
Second government deployment
A separate isolated contour for the avatar platform, deployed on customer infrastructure by the same team from the same codebase.
Our own AIvatars platform
A multi-tenant SaaS with autoscaling GPU pools, running alongside the on-prem installations from the same codebase.
Research heritage
Rooted in research.
Selected peer-reviewed publications by team members.
- Large Residual Multiple View 3D CNN for False Positive Reduction in Pulmonary Nodule Detection — IEEE CIBCB 2017, Manchester. DOI 10.1109/CIBCB.2017.8058549
- 2018 Robotic Scene Segmentation Challenge — MICCAI 2018 EndoVis Workshop / arXiv 2001.11190 (collaborative)
Further technical reports cover BI-RADS mammography classification, aorta and annulus modelling for heart-valve sizing on contrast CT, and denoising auto-encoders for bone suppression. Full publication list and named CVs available under NDA.
The people
Built by a team that ships.
A small group splitting time between engineering, design, and customer pilots.
In the wild
Avatars on real screens.
Photos and clips from kiosks, hardware tests, and pilot deployments.
Let us talk about your project
We are happy to share architecture diagrams and recordings from live deployments, provide named CVs per role, and discuss engagement shape — from a short design audit to turnkey delivery.