I am an Autonomous Systems R&D Engineer engineering software architectures where real-time determinism, low latency, and robust physical safety are paramount. My core research and development work bridges autonomous vehicles (Autoware / AWSIM), aerial defense & UAV interception systems (PX4 / ROS 2), and high-bandwidth digital twin telemetry interfaces (React Three Fiber / Three.js / WebSockets).

• Autonomous Driving & Ground Mobility: Autoware.Universe software stack, AWSIM digital twin simulations, Lanelet2 HD map processing, ego-trajectory generation, and in-vehicle 3D digital twin cockpit visualization.
• Aerial Autonomy & Counter-UAS: Optical-seeker kinetic anti-drone systems (pure vision-based 3D tracking & terminal sprint), PX4 Autopilot SITL, MAVLink telemetry pipelines, custom Ground Control Stations (BagdarGC / QGroundControl), and competitive dogfight autonomy (TEKNOFEST Savaşan İHA).
• Edge Perception & Real-Time Computing: Real-time Linux (PREEMPT_RT), zero-copy DDS middleware (CycloneDDS), NVIDIA Jetson edge compute (TensorRT FP16 / CUDA), LiDAR/Camera extrinsic calibration, and multi-threaded C++20 nodes.
• Modern Web & Mobile Product Ecosystems: End-to-end full-stack architectures, award-winning agricultural platforms (TEKNOFEST 1st Place), NestJS microservices, React Native mobile clients, and high-frequency financial dashboards.

Distinct application-layer toolchain engineered for in-vehicle 3D HMI cockpits, responsive dashboards, and cross-platform mobile ecosystems.

Layer	Technologies & Frameworks
Autonomous Driving
UAV Autonomy & Defense
Robotics Middleware & Core
Perception & Edge Compute
OS & Execution Runtime

Real-time autonomy perception-to-cockpit visualization pipeline interfacing with Autoware.Universe and digital twin simulations.

• In-Vehicle 3D Cockpit: Built autoware-hmi-visualization using React Three Fiber, Three.js, and TypeScript to render ego-vehicle dynamics, Lanelet2 vector roads, lane markings, surrounding environment blocks, and planned trajectory ribbons in real-time.
• Bridge Architecture: Designed a low-latency pipeline connecting Autoware Odometry, Trajectory, and Lanelet2 MapBin topics through a C++ map adapter and a Python ROS 2 / TF bridge transmitting normalized JSON over WebSockets.
• Simulation Validation: Integrated testing workflows connecting Autoware with AWSIM digital twins and Gazebo environments for closed-loop trajectory following.

Autonomous hit-to-kill interception system operating on pure visual feedback without ground-truth coordinate cheats.

• Optical Seeker Guidance: Integrated a 20° elevated seeker camera on a custom PX4 airframe in Gazebo. Utilized OpenCV HSV masking and contour moments to extract tracking errors ($e_x, e_y$) and convert them into 3D Line-of-Sight (LOS) rate commands.
• Breadcrumb Dead Reckoning & Terminal Ramming: Designed a coordinate breadcrumb memory system; if the target executes an evasive maneuver outside FOV, the interceptor maneuvers towards the last calculated spatial trajectory, reacquires visual lock, and initiates a terminal sprint at 25+ m/s (90+ km/h) for kinetic neutralization.
• Simulation Bridge: Developed a high-frequency C++ bridge (gz-transport to OpenCV) processing live camera feeds and contact sensor telemetry at sub-millisecond cycles.

Autonomy software suite for competitive multi-UAV mission management and dogfighting.

• Mission Execution & Visual Lock: Developed state machine sequencers handling autonomous takeoff, waypoint routing, computer-vision target detection (YOLO/OpenCV), and active target tracking over MAVLink/Pixhawk.
• Custom Ground Control Stations: Engineered custom GCS tooling (VaryansGC & BagdarGC) based on QGroundControl architecture for real-time mission telemetry, command injection, and fail-safe oversight.
• Repositories: varyans · varyans_workflow · VaryansGC · Teknofest-SIHA-Goruntu-Isleme-Gorev-Yazilimi

Award-winning full-stack agricultural ecosystem connecting producers, consumers, and supply chain administrators.

• Backend Architecture: Robust NestJS REST API with JWT authentication, RBAC authorization, Swagger OpenAPI documentation, and comprehensive automated test suites (Jest/Supertest).
• Data & Caching: PostgreSQL persistence layer managed with Prisma ORM, automated migrations, Redis caching, and Docker Compose orchestration.
• Mobile Experience: Cross-platform React Native / Expo application with Redux Toolkit state management and offline-first synchronization.
• Repositories: FarmerAIMobileApp · RestApiFarmerai · FarmerAIOCR

Perception and robotics experimentation packages for multi-modal sensor suites.

• LiDAR-Camera Calibration: tetra_calibration — Extrinsic calibration pipeline aligning 3D LiDAR point clouds with 2D optical cameras for fused spatial projection.
• Autonomous Simulation: Tetra_Otonom — ROS Noetic / Gazebo simulation implementing spatial obstacle avoidance and LiDAR-driven state machines.

Interactive multi-panel financial market analytics interface.

• Architecture: Next.js App Router, TypeScript, and Tailwind CSS.
• Performance: High-frequency chart controls, validated input parameters, and modular UI components engineered for minimal re-renders.
• Repository: trading-viewer

   [ Deterministic ]       [ Simulation-First ]       [ Hardware-Conscious ]
          │                         │                          │
          ▼                         ▼                          ▼
 Loops hold period         Validating QoS & jitter      Bounded memory & zero-copy
 under worst-case load     before burning flight hours   buffers on Jetson / ARM

• Deterministic Real-Time Discipline: Prioritizing predictable loop frequencies (SCHED_FIFO, PREEMPT_RT), bounded heap allocations in real-time control paths, and explicit fail-safe state machines over unvalidated abstractions.
• Simulation-First, Hardware-Proven: Validating sensor degradation, DDS QoS latency, and edge cases in digital twins (AWSIM, Gazebo Harmonic, SITL) before conducting flight tests or physical vehicle trials.
• Hardware-Conscious Edge AI: Optimizing neural network inference via TensorRT FP16, implementing zero-copy pipelines, and respecting memory bandwidth and thermal limits on NVIDIA Jetson SoCs.
• End-to-End Architectural Integrity: Bridging low-level C++ drivers and robotics middleware seamlessly with high-bandwidth 3D visualization and cloud/edge interfaces.