Hardware that measures what insurers can price.
I'm Yacine. I build Datum Industries — instrumented protective equipment that turns how warehouse workers move into a risk signal their insurer will pay for.
Backs give out before machines do.
Musculoskeletal injury is the largest preventable cost in physical logistics. It is measured once a year, by survey, after the damage is done. Nobody knows which shift, which task, or which worker produced it.
Source: HSE, Health and Safety at Work — Summary Statistics for Great Britain, 2025.
A wearable that reports posture, load, and fatigue in real time.
Three instrumented pieces of kit that replace equipment workers already wear, so adoption costs an operator nothing in behaviour change. Full hardware and software stack built solo.
Instrumented glove
Paired ICM-42688-P IMUs on a shared clock for common-mode rejection, force-sensing resistor, capacitive engagement electrode. 120 Hz production sampling.
Lumbar jacket
Spinal flexion and load estimation across the lift cycle — the movement that produces most of the injury cost.
Safety eyewear
Gaze context, so a risky posture can be attributed to the task that caused it rather than the worker performing it.
Telematics, for bodies.
Black-box insurance worked because a driver would trade data for a cheaper premium. The same trade exists in logistics and nobody has built the box. Datum sells verified injury reduction to the insurer, and the insurer discounts the operator's premium — so the operator's cost of adoption is negative.
Movement data from equipment already required on site.
Per-task and per-site musculoskeletal risk, updated continuously.
Underwriters get a live loss signal instead of an annual claims history.
Premium reduction flows back to the operator. Injury rate falls.
Validated precedent in the US market. Entry route in the UK is via MGAs and specialist brokers rather than large carriers — shorter decision chains, real appetite for differentiated risk.
The wearable funds the chip.
Datum is a data business first, but the reason it exists is what sits underneath it. I hold a pending patent in photonic AI inference hardware — inference at a fraction of the power budget of an equivalent electronic part, with inherent radiation tolerance.
Those two properties matter almost nowhere on Earth and matter enormously off it. Autonomous robotics in orbit and on the lunar surface need on-board inference that survives the radiation environment and fits inside a power budget measured in watts. The wearable platform generates the revenue, the training data, and the deployment discipline to get there.
I build the whole stack, and I finish things.
Mechanical engineering at Queen Mary University of London. I designed the sensor hardware, wrote the firmware, built the host-side pipeline, and ran the market validation myself — not because that's ideal, but because it's what the stage required.
RSS Fellowship
Recognition for research contribution.
Patent pending
GB 2607326.2, photonic AI inference hardware.
BEng Mechanical Engineering
Queen Mary University of London, 2026.
Full-stack build
Sensor firmware in embedded C++, binary wire protocol, host analysis toolkit, ML training pipeline.
Open to conversations with investors, insurers, and logistics operators who want to run a pilot.