Key Skills
Industry
Job Description
Company: Kasava AI Pvt Ltd (engineering hub for Kasava, UK)
Location: New Town, Kolkata
Reports to: Founder & CEO
Type: Full-time
About Kasava
Kasava builds cryptographic trust infrastructure for fleet insurance. We turn dashcam footage and telematics into tamper-evident, court-ready evidence, so insurers and fleet operators can automate claims, cut fraud, and settle disputes on proof rather than argument. Our patent-pending technology detects incidents and signs the evidence at the point of capture, then maintains a verifiable chain of custody all the way to the insurer. UK business, tight engineering hub in Kolkata, fast pace, real product in the field
The Role
You’ll own the on-device software runtime on our Linux dashcam platform, the layer that makes everything actually run on the hardware, in real time, in a moving vehicle. You take our computer-vision detection models and our cryptographic signing from the lab onto resource-constrained edge devices and make them reliable in the field.
This is hard, real engineering: real-time inference and signing co-existing on constrained hardware, surviving firmware updates, working under field conditions. You’ll partner with our CV engineer (who builds the models) and our cryptography architect (who designs the signing scheme), your job is to make it all run, fast and dependably, on the device
What You’ll Do
• Build, optimise, and deploy real-time computer-vision inference on our Linux edge devices (TensorRT / ONNX, quantisation, low-latency pipelines).
• Implement and run the on-device signing under the cryptography architect’s design, including secure handling of keys on the hardware.
• Own the device runtime: system services, resource management, reliability, and persistence of our software across over-the-air updates.
• Integrate camera and sensor data and video pipelines on-device.
• Work at shell level on the device to run and maintain our proprietary stack on the platform.
• Partner with the CV, backend, and crypto teams on the end-to-end pipeline from capture to signed packet
What We’re Looking For (Essential)
• Strong embedded / systems Linux: building, deploying, and running software on resource-constrained Linux devices, Linux internals, system services, cross-compilation.
• Solid C/C++ (Python useful day to day; Rust a plus).
• Real-time / low-latency on-device work and genuine performance optimisation.
• Edge AI deployment: getting ML / CV models running on edge hardware — TensorRT, ONNX, quantisation, or equivalent.
• The pragmatism to make things work reliably on imperfect hardware in the field
Nice to Have
• On-device security: secure key storage, TEE / TrustZone, secure boot (you’ll partner with our crypto architect — you implement, they design).
• Video pipelines (GStreamer, RTSP) and camera / sensor integration.
• Automotive, IoT, or device / firmware background.
• CUDA or other hardware-accelerated inference.
Why This Role
You own the layer that turns models and cryptography into something that actually runs in a moving vehicle, the part most engineers never get to touch. Small team, fast shipping, real ownership. If you like edge systems where the constraints are real and the work ships to the field, get in touch.
