TJL Defence
Defence intelligence, from the first sensor to the last decision.
Purpose-built for operations that cannot fail, and designed, built and run in the Kingdom.
Who we are
A Saudi deep-tech company that makes the unseen visible.
TJL takes its name from تَجَلّى: to manifest, to become visible. One algorithmic engine runs across energy, mobility and defence, and what it learns in one domain strengthens the others.
One algorithmic engine · every domain
TJL Defence
From sensing to action, one chain.
Each works on its own and alongside the systems already in service. Together they connect every unit, see and predict, understand every link, and coordinate the fleet.
Communication infrastructure
One sovereign mesh. Nobody external can switch it off.
TJL owns and operates the mesh inside the Kingdom, independent of commercial networks. Defence units run their own missions on it, and it heals itself around any lost node.
- No SIMNo commercial operator anywhere in the chain
- Self-healingTraffic finds a new route the moment a node is lost
- Isolated by keysEach unit's data can be read only with its own keys
- Multi-purposeSensors, trackers and control on battery-powered devices, on one mesh
Illustration with example values
Communication infrastructure
Many units on one mesh. Each sees only its own.
Every unit has its own keys, its own data and its own API. TJL Vectora, a mini C4 platform, gives each one its own operating picture and alerts.
Illustration with example values
Communication infrastructure
Coverage is the capability, built base by base, corridor by corridor.
Fixed units are served from the first installation, which justifies the buildout. Mobile units are fully served as the footprint grows.
Illustration with example values
Illustration with example values
- MeasureDistance, direction and speed for every target
- PredictA short-horizon trajectory for each tracked target
- DecideThreat level from where a target is heading, not only where it is
- OfflineAll on the device: no cloud, no connectivity needed
Trajectory prediction
See it, track it, and know where it is going next.
TJL Intelligence detects and tracks targets with a camera on an embedded board. Each target's distance, direction and speed are measured and its path predicted, on the device.
- Objects
- 0
- Nearest
- ·
- Distance
- ·
- Speed
- ·
- Heading
- ·
- Threat
- ·
Trajectory prediction
All the intelligence on the device.
From the vision sensor to the response, every step runs on one embedded processor. A reasoning layer then answers questions about each session.
- No cloudEvery frame is processed where it is captured
- IntegratesWorks with existing cameras, displays and systems
- ExplainsQuestions about a session are answered from its own records
Most targets came from the left: 7 of 12 between −60° and −30°. The nearest approach was 18 m.
Illustration with example values
- 1A closed link
- 2TJL listens and learns
- 3Understood in C2
3c 00 05 06 b8 00 03Track 07 · 12.4 km · bearing 214° · closingIllustration with example values
Communication decoding
Closed protocols, decoded by AI agents from their own traffic.
Legacy and foreign-vendor systems often speak a language only their own consoles understand. TJL listens to the link, works the language out, and brings every message into national C2. Nothing is sent to either device.
Communication decoding
Every closed link becomes a source you control.
The output is a decoder with its tests and its specification, not a model in the packet path. It can read the link, and it can speak it.
- DeterministicPlain code runs on the traffic; no AI sits in the packet path
- Listen-onlyNothing is sent on the link, so the systems on it keep running exactly as they are
- Vendor-independentIntegration no longer waits on the original vendor or its documentation
- CumulativeEach decoded protocol joins a library, so the next one is faster
- Listen
- Structure
- Decode by difference
- Validate
- Run live
A lead agent proposes each finding. A second agent, with fresh context, tries to break it before it counts.
A simulation running in this browser on an example scenario: the plan, the paths and every replan are computed live.
UAV mission engine
Intelligence above the fleet, not inside each drone.
Mission goals become tasks. Each task goes to the right UAV by capability and position, every path is flyable, and the plan changes the moment a UAV is lost or new orders arrive.
UAV mission engine
Four layers. One chain of command.
The engine sits between the operator and the fleet, and drops into existing C2 through its API. No rip-and-replace.
UAV mission engine
Five mission profiles. All high stakes.
From a small tactical team to a full swarm, with encrypted links and operations that continue as communications degrade.
Illustration with example values
TJL Defence · Together
From detection to response, one operating picture.
An example at a perimeter at night. Each step is a different TJL capability, and the operator sees all of them in one place.
- 1
Detect
TJL Intelligence picks up a target at the perimeter camera.
- 2
Predict
Its path is predicted to cross the fence line.
- 3
Alert
The alert crosses the sovereign mesh to the perimeter unit, under its own keys.
- 4
Confirm
A legacy radar, decoded by TJL, confirms the track.
- 5
Re-task
The mission engine sends the two nearest UAVs to the predicted crossing.
- 6
Decide
The operator sees the target, its prediction and the UAVs in one picture.
Connect. Sense. Understand. Act.
TJL Defence: built in the Kingdom, for operations that cannot fail.