Senior Computer Vision Engineer – UAS Tracking

Ground A; Berlin; Vor Ort; EUR 90.000 - 130.000; Vollzeit

Benefits dieser Stelle

VSOP / equity participation

Central Berlin office

Extreme growth opportunity in a fast-s

Monthly voucher

Public transportation subsidies

Zusammenfassung

Ground A, a defense technology startup in Berlin, seeks a Senior Computer Vision Engineer – UAS Tracking to design and implement the tracking layer of our perception stack for multi-UAS using IR/thermal data, fused with laser range finder, radar, and IMU measurements. The role emphasizes real-time tracking with low latency, robust state estimation, and continual improvement under operational conditions.

The position is on-site in Berlin, requiring NATO member state citizenship.

Qualifikationen

• 6+ years of experience in computer vision, visual tracking, estimation, or real-world perception systems.
• Strong hands-on experience with multi-object tracking or visual target tracking.
• Practical experience with tracking-by-detection pipelines.
• Strong understanding of estimation methods, especially Kalman filters, EKF, UKF, or related approaches.
• Experience deploying perception or tracking systems on edge platforms, ideally NVIDIA Jetson / Tegra.
• Experience with TensorRT and/or CUDA.

Aufgaben

• Design, implement, and improve multi-UAS tracking pipelines for real-world aerial targets.
• Develop and maintain tracking-by-detection systems using detector outputs as inputs.
• Track initiation, confirmation, continuity, and lifecycle management.
• Track confidence scoring and loss handling with reacquisition.
• Model measurement uncertainty and propagate confidence into downstream systems.
• Develop state estimation pipelines for position, velocity, trajectory, and uncertainty.
• Apply filtering approaches like Kalman/Extended/Unscented Kalman filters.

Kenntnisse

Computer vision

Visual tracking

Estimation methods

Kalman filters/EKF/UKF

Tracking-by-detection

Edge deployment

C++ production

Python

NVIDIA Jetson/Tegra

TensorRT/CUDA

Tools

NVIDIA Jetson / Tegra

TensorRT

CUDA

Jobbeschreibung

Location: Berlin (Onsite)

About Us

We are a defense technology startup based in Berlin, building the next generation defense layer against aerial threats. Our team develops mission-critical systems that combine hardware, software, sensing, and real-time AI to operate reliably in complex real-world environments.

Operating in stealth mode, we place strong emphasis on engineering rigor, real-world performance, and defense-grade system reliability.

Role Overview

We are seeking a Senior Computer Vision Engineer – UAS Tracking to design, implement, and continuously improve the tracking layer of our perception stack.

This role is focused on multi-UAS tracking using primarily IR / thermal camera data, combined with additional measurements from sensors such as laser range finders, radar, and IMUs. The core challenge is not simply detecting a drone in a single frame — it is maintaining stable, low-latency, confidence-scored tracks over time in difficult real-world conditions.

You will consume detector outputs and develop robust tracking pipelines that support real-time target following, aiming, trajectory prediction, and downstream control. Your work will directly influence how reliably the system can maintain track continuity, reject false tracks, reacquire targets after occlusion, and provide actionable target state estimates under operational constraints.

This is a hands‑on, production‑oriented role for someone with deep experience in visual tracking, estimation, and real‑time perception systems.

This position requires onsite work in Berlin and candidates must be from a NATO member state.

Key Responsibilities

Multi‑UAS Tracking & Track Management

• Design, implement, and improve multi-UAS tracking pipelines for real-world aerial targets
• Develop and maintain tracking-by-detection systems using detector outputs as inputs
• Track initiation and confirmation
• Track continuity and lifecycle management
• Track confidence scoring
• Track loss handling and reacquisition
• Multi-target separation and identity consistency
• Reduce track fragmentation, identity switches, jitter, and unstable target handoff
• Define and track performance metrics such as: Track continuity, Latency, ID switches, False tracks, Missed tracks, Reacquisition time, Prediction stability
• Develop and tune state estimation pipelines for target position, velocity, trajectory, and uncertainty
• Apply and extend filtering approaches such as Kalman filters, Extended Kalman filters, Unscented Kalman filters, and related Bayesian estimation methods
• Design prediction logic for short‑horizon target motion and aiming support
• Improve robustness under noisy detections, intermittent measurements, and sensor dropouts
• Model measurement uncertainty and propagate confidence into downstream systems

IR / Thermal Vision Tracking

• Develop tracking approaches optimized for IR / thermal camera data
• Address challenges such as low target pixel footprint, weak contrast, thermal clutter, atmospheric effects, motion blur, sensor noise and artifacts, and camera movement and vibration
• Improve tracking robustness under poor visibility, weather variation, and changing environmental conditions
• Use feature tracking, temporal association, and re‑identification approaches where appropriate

Sensor Fusion for Tracking

• Integrate visual tracks with additional measurement sources, including laser range finder data, radar measurements, and IMU data
• Fuse heterogeneous measurements to improve track stability, target state estimation, and trajectory prediction
• Collaborate with robotics and controls engineers to ensure tracking outputs are low‑latency, stable, predictive, confidence‑scored, and suitable for real‑time aiming and following

Real‑Time Deployment & Optimization

• Deploy and optimize tracking pipelines on NVIDIA Jetson / Tegra platforms
• Optimize performance using TensorRT, CUDA, efficient video and preprocessing pipelines, and production‑grade C++ integration
• Balance tracking quality, latency, memory usage, and thermal constraints
• Ensure tracking pipelines operate reliably within real‑time system constraints

Evaluation, Field Data & Continuous Improvement

• Analyze real‑world tracking failures and translate them into algorithmic improvements
• Build evaluation datasets and metrics for multi‑UAS tracking performance
• Support field testing and post‑test analysis
• Collaborate with engineering teams to improve robustness based on operational data
• Maintain clear documentation of tracking assumptions, models, parameters, and performance limits

Required Qualifications

• 6+ years of experience in computer vision, visual tracking, estimation, or real‑world perception systems
• Strong hands‑on experience with multi‑object tracking or visual target tracking
• Practical experience with tracking‑by‑detection pipelines
• Strong understanding of estimation methods, especially Kalman filters, EKF, UKF, or related approaches
• Experience with feature tracking, temporal association, re‑identification, or trajectory prediction
• Experience working with real‑world video data, including noisy measurements, sensor artifacts, domain shift, and difficult edge cases
• Strong Python skills and production‑level C++ experience
• Experience deploying perception or tracking systems on edge platforms, ideally NVIDIA Jetson / Tegra
• Experience with TensorRT and/or CUDA
• Ability to work closely with robotics, controls, and systems engineers
• Onsite availability in Berlin
• Citizenship of a NATO member state

Preferred Qualifications

• Background in defense, aerospace, robotics, autonomous systems, or industrial perception
• Experience with IR / thermal tracking
• Experience with long‑range small‑object tracking
• Experience fusing visual tracks with LRF, radar, or IMU measurements
• Experience with radar‑assisted or range‑assisted target state estimation
• Experience building real‑time perception‑to‑control pipelines
• Experience supporting field tests or deployed perception systems

What We Offer

• VSOP / equity participation
• Central Berlin office
• Extreme growth opportunity in a fast‑scaling startup
• Monthly voucher
• Public transportation subsidies

Who You Are

• Tracking‑first: You think in tracks, trajectories, uncertainty, continuity, and reacquisition
• Estimation‑driven: You understand that robust tracking requires more than frame‑by‑frame detections
• Real‑world focused: You care about performance under noise, clutter, occlusion, vibration, and weather
• Low‑latency oriented: You understand that tracking quality only matters if it arrives on time
• Analytical: You debug failure modes with data, models, and clear metrics
• Collaborative: You work effectively with robotics and controls engineers
• Mission‑aligned: You understand the responsibility of building perception systems for defense applications