Technical scope

Full Tech Stack & Technical Expertise

Below is a breakdown of the technologies, standards and tools I work with on production autonomous systems: from PCBs and DO-178C-class flight firmware, through navigation and SLAM, edge AI, to cloud-native architecture. This is a real technology stack, applied in practice among other things while building my own inspection drone platform, where I own the full technical architecture as Founder/CTO.

Hardware & Electronics

Full-cycle electronics design for autonomous platforms.

MicrocontrollersARM Cortex-M (ST, TI, Nordic, Silicon Labs), Cortex-A (ST, NXP), NVIDIA Jetson.

Sensors & positioningIMU, magnetometer, barometer, redundancy, RTK GNSS.

PerceptionRGB cameras, thermal cameras, LiDAR.

BusesEthernet/DDS, DroneCAN, UAVCAN, I2C, SPI, UART.

PCB & productionAltium Designer, SMT assembly (JLCPCB, PCBWay), ENIG finish.

Firmware & Flight Control

Safety-critical software, rigor close to aviation standards.

RTOS & languageZephyr, FreeRTOS, C++17, static allocation only.

ComplianceRigor close to DO-178C DAL-C, MC/DC, FMEA, WCET.

State estimationEKF, UKF, MEKF, IEKF, FIR/IIR.

ControlCascaded PID, anti-windup, motor mixer.

SafetySeparation from the mission computer, automatic RTH.

Navigation & SLAM

Precision navigation with GNSS and in GPS-denied environments.

Sensor fusionAHRS, GPS/INS, quaternions, sensor redundancy.

RTK headingMoving-baseline, independent of the magnetometer.

SLAMFAST-LIO2, ORB-SLAM3, RTAB-Map, PCL, GTSAM, Ceres.

PlatformROS/ROS2, Gazebo, Jetson-class computers.

Edge Computing & Artificial Intelligence

Perception models and a deployment pipeline for edge and cloud.

FrameworkROS2, OpenCV, CUDA, gRPC/REST.

Model trainingPyTorch, TensorFlow, YOLOv8, MLflow, Weights & Biases.

Domain modelsDefect detection and segmentation: YOLOv8, ResNet, PointNet++.

DeploymentONNX, TensorRT, Triton, FP16/INT8 quantization.

MLOpsHuman-in-the-Loop, model versioning.

Protocols, API & Integration

Contract-first: one contract, many languages.

Data contractsProtobuf, multi-language SDKs (C, C++, Python, TypeScript).

ProtocolsMAVLink, DroneCAN, UAVCAN, DDS, gRPC, REST, OPC-UA, MQTT.

ConnectivityLow-latency radio, LTE, BLE, Zigbee, Sub-G as backup.

SecuritymTLS, OAuth2, JWT, TLS/mbedTLS.

Backend & Cloud Infrastructure

Cloud microservices, split into core and domain services.

Application stackPython, FastAPI, PostgreSQL, Redis, Celery.

InfrastructureGCP, AWS, Azure, serverless.

IaCTerraform, Docker, Kubernetes.

CI/CDGitHub Actions, backups, multi-region replication.

Frontend & Desktop Applications

Web dashboards and a Ground Control Station-class desktop app.

WebReact, Next.js, TypeScript, Tailwind CSS.

Desktop (GCS)Qt/QML, C++17, MVVM, gRPC.

FeaturesMission planning, controller tuning, diagnostics, flight logs.

Simulation, Testing & DevOps

The same code on hardware and in simulation, before it reaches production.

SimulationGazebo, SiL/HiL, native firmware simulation.

End-to-end testingRobot Framework, pytest, smoke/nightly/soak.

Code qualityclang-format, clang-tidy, quality gates.

CI/CDGitHub Actions, Jenkins, Gerrit.

Your team works on a different stack, or you need a review of a specific part of the architecture? Let's book a 30-minute diagnostic call.

Book a call Visit the Blog