Our client is a European deep-tech startup developing next-generation radar systems for the detection and tracking of small, low-flying drones.
The team works across radar architecture, RF hardware, signal processing, embedded systems and data-processing software. Engineers have real ownership of technical decisions and a direct impact on system performance and product capabilities.
This is a hands-on R&D environment for specialists who want to solve complex engineering problems and contribute to defence technology designed for real-world deployment.
What you'll do
- Pipeline — Build a Python backend that ingests a live I/Q stream and manages queues, buffers, processes and data flow.
- Real-time path — Move large I/Q matrices into the AI core with minimal latency using shared memory, zero-copy and ring / ping-pong buffers — not REST.
- Integration — Wrap and profile the team's math/AI modules (NumPy, ONNX Runtime / PyTorch) and remove bottlenecks together with the signal engineers.
- Data layer — Design the schema for tracks, events and metadata; async writes via asyncpg / SQLAlchemy async. Raw I/Q stays in RAM; raw logs go to cold storage.
- API & UI — FastAPI for commands, auth and status; secure WebSockets streaming tracks to a simple operator dashboard (Plotly Dash / Streamlit).
- Infra — Docker / Docker Compose, GitHub Actions CI/CD, shared-memory config in containers, Terraform-ready deployment.
- Security — TLS to Postgres, HTTPS/WSS, JWT auth, network isolation, secrets handling, least-privilege access.
What we're looking for
- Senior (or strong mid) Python engineer who has built real engineering systems, not only web APIs.
- Solid with asyncio, multiprocessing and multiprocessing.shared_memory.
- Comfortable with large data matrices (NumPy) and profiling / optimising Python for latency and throughput.
- PostgreSQL (asyncpg or SQLAlchemy async), FastAPI, REST, WebSockets.
- Docker, Docker Compose, Linux, Git / GitHub Actions.
- Security basics: TLS, HTTPS, WSS, JWT.
Nice to have
- ONNX Runtime / PyTorch inference, CuPy, Numba, SciPy, GPU inference.
- ZeroMQ / PyZMQ, Zarr / HDF5, structured logging, Prometheus / Grafana.
- Linux performance tuning (CPU affinity, taskset, psutil), Terraform, Ansible.
- Any exposure to high-throughput / low-latency systems, SDR, radar, signal processing, streaming or telemetry pipelines.
WHAT WE OFFER
Impact
Own the core signal-detection system. Your work directly shapes what the product can detect, understand, and act on.
Engineering Challenges Worth Solving
Work on complex radar, DSP, and signal-processing problems that are typically found only in large R&D organizations or deep-tech companies. You'll tackle hard technical challenges with direct influence on outcomes.
A Team You'll Learn From
Collaborate closely with scientists, engineers, and operators who combine deep technical expertise with a strong product mindset. No layers of bureaucracy—just smart people solving meaningful problems together.
Ownership & Autonomy
We're looking for builders, not executors. You'll have the freedom to make decisions, take responsibility, and drive critical parts of the technology forward.