Bridging the gap between Embedded Hardware, Machine Learning, and Secure Infrastructure.
I am an Automation & Robotics Engineer (M.Sc.) obsessed with the convergence of hardware and software. I don't just write code; I build ecosystems.
My expertise spans the full stack of physical computing: from designing custom PCBs and writing firmware, to deploying ML models at the edge and managing cloud infrastructure.
I bridge the gap between "it works on my machine" and "it works in the field." I am currently seeking complex engineering challenges in Software Development and MLOps.
Automation & Robotics
University of Zielona Góra
Automation & Robotics
University of Zielona Góra
EE.02 & EE.21 Quals
CKZiU Nr.2 "Elektronik"
A holistic engineering approach: From identifying the problem to delivering the physical and digital solution.
Ekoenergetyka Service
03.2025 — Present
Managing telemetry and diagnostics for high-Power EV charging infrastructure.
Calmet Sp. z o.o.
MPLAB Robotics
Association of Polish Electrical Engineers
Authorized for operation, maintenance, and measurement of electrical installations up to 1kV.
Siemens Academy
Advanced training in SIMATIC S7 controllers and industrial automation systems logic.
Selected works in Robotics, ML Systems, and Infrastructure.
A distributed Neural Architecture Search (NAS) system that orchestrates GPU training nodes and visualizes 3D network topologies via WebSockets. It benchmarks TFLite models on Raspberry Pi edge devices using a custom raw-TCP protocol.
Features a custom latency-aware objective function (MSE + LatencyPenalty) to optimize for real-time edge performance, utilizing Optuna pruning to terminate inefficient architectures early.
A centralized IDP architecture using Keycloak and Nginx. Implements Identity Propagation to inject user headers into backend apps, enabling Just-In-Time (JIT) provisioning without local password databases.
Architected as a Split-Stack Hybrid Cloud: A public VPS acts as the secure ingress gateway, tunneling heavy traffic to a private home lab via a monitored OpenVPN split-tunnel network.
A real-time object detection system running on Raspberry Pi 5 + Hailo-8L. Trained a custom YOLOv8 model on a manually annotated dataset of 40+ electronic components (capacitors, batteries, modules).
Engineered the full compilation pipeline: Quantization, conversion to ONNX, and compilation to HEF (Hailo Executable Format) for hardware acceleration.
A planar robotic arm for hazardous liquid distribution. Powered by STM32F303 running FreeRTOS. Features a custom C++ Inverse Kinematics (IK) solver and a 3D-printed differential drive mechanism designed in Fusion 360.
High-performance remote built on STM32H7. Runs a 60fps TouchGFX interface on FreeRTOS. Handles 5 analog channels, bidirectional UART/XBee telemetry, and context switching for multiple vehicle profiles.
A multi-user REST API with a Split-Stack architecture (Public VPS frontend, Private Homeserver DB). Relies on Nginx/Keycloak for auth and uses HTTP Headers for JIT user provisioning.
Custom router built on Raspberry Pi 5 with PCIe 4G modem. Implements `mwan3` for load balancing between Landline and LTE. Features a custom SMS control daemon and automated PCIe device recovery scripts.
A Python-based orchestrator for creating isolated, reproducible Docker environments (Sandbox/CTF). Generates Dockerfiles from JSON config, handles SSH keys, and manages container lifecycle via an interactive CLI.
A lightweight graphics engine written in C that renders 3D .obj models directly in the terminal using ncurses. Features a custom rasterization pipeline, Z-buffering, particle effects, and billboard text support.
A hybrid-cloud environment where the Public VPS handles security and ingress, while heavy compute and ML tasks are tunneled to the Private Home Lab.
The Public Entrypoint
Ubuntu • ML Worker
Raspberry Pi 5
Rapid Prototyping