What if your code could give someone a voice?
At Signapse, we’re building real-time AI that translates spoken and written language into sign language video making transport, healthcare, and digital life accessible to Deaf communities around the world. We’re a small, fast-moving team and we’re looking for an ML Systems Engineer to help us scale.
“You’ll be at the forefront of one of the first real-time AI sign language generation systems in the world.”
About Signapse
Signapse is a fast-paced tech-for-good start up on a mission to make the world more inclusive for the Deaf community. Using AI, we create sign language translations across video, transportation, and web. We are now looking to expand our exceptional team.
We are able to consider candidates with location rights to work from priorities locations with rights to work in the UK or EU. English spoken and written must be fluent.
The Challenge
Our generative AI models produce sign language video in real time, delivered over GPU infrastructure across cloud and on-prem environments to a global audience. The engineering problem is genuinely hard: keep latency low, maximise GPU utilisation, and build infrastructure that scales to hundreds of simultaneous streams. You’ll work across the full ML inference stack from model optimisation to deployment infrastructure and own this challenge.
What you’ll work on
ML inference optimisation
* Profile and optimise deep learning models used for sign language video generation
* Reduce inference latency using quantisation, pruning, mixed precision, and kernel optimisation
* Improve GPU utilisation and throughput across inference pipelines
* Work closely with ML researchers to ensure models are production-ready
ML infrastructure & deployment
* Build and maintain scalable model serving systems on GPU clusters
* Design autoscaling infrastructure to meet real-time SLAs
* Contribute to model deployment pipelines, versioning, and rollback strategies
Performance engineering
* Develop benchmarking frameworks for tracking inference performance
* Identify bottlenecks across the ML pipeline and eliminate latency hotspots
* Implement performance monitoring and alerting for production systems
* Evaluate new hardware accelerators and inference run times.
Scaling globally
* Work with the research team to expand sign languages and digital signers
* Architect systems that allow rapid onboarding of new languages
* Build low-latency infrastructure that scales to hundreds of concurrent streams
What we’re looking for
Essential
* 3+ years of experience in ML systems engineering, ML infrastructure, or backend systems.
* Strong Python skills (Rust is a bonus)
* Experience working with production ML models
* Strong debugging, profiling, and performance analysis skills
* A genuine interest in building latency-critical, high-throughput systems
Desirable
* Experience with TensorRT, ONNX, Triton, TorchServe, or similar inference tools.
* Familiarity with GPU architecture and performance optimisation
* Experience with video, graphics, or real-time streaming systems (HLS, RTMP,
* SRT)Experience with Kubernetes, Docker, and ML workloads at scale.
* Familiarity with AWS, including Sage Maker.
Our current stack
Python, Go, Rust · PyTorch · Kubernetes (cloud + on-prem) · AWS + SageMaker · Real-time streaming (HLS, LL-HLS, RTMP, SRT) · GPU inference workloads.
Why join Signapse.
* Work on technology that directly improves accessibility for Deaf communities worldwide
* Help build one of the first real-time AI sign language generation systems in existence
* Join a small, experienced engineering team solving genuinely hard technical problems
* Take real ownership of critical systems as an early engineering hire
* 24 days’ holiday + bank holidays, company pension, and free sign language classes
* Remote-first with optional London days.
Our commitment to inclusion:
Signapse is committed to building a team that is as diverse as the communities we serve. Any qualified applicants who are native sign language users are guaranteed an interview.