Tomáš Buša

Experience

Thesis research at Imperial, and three summers at MicroStep in Bratislava.

2025 – 2026

MSc thesis research, Imperial College London

AccNet, supervised by Dr Marios Kogias; co-developed with a co-author of Machnet

  • Extended a custom Linux kernel's system-call path with generic entry and exit eBPF hooks that can rewrite an intercepted system call into a different one, or implement and short-circuit it entirely.
  • Designed and built the interception layer for a transparent network-stack bypass: cgroup eBPF hooks redirect the data of unmodified TCP applications through shared-memory ring buffers to Machnet, a DPDK-based transport.
  • Added two kernel modifications, a writeable ring-buffer mapping and eventfd kfuncs, to get a zero-copy transmit path and kernel-style wake-ups without touching the application.
  • Evaluated against tuned Linux TCP with memcached, lighttpd and an unmodified Go server on CloudLab hardware.
Jun – Aug 2024

Software engineer intern, MicroStep

Bratislava, Slovakia

  • Built an automation tool that parses electrical schematics of CNC machines and generates YAML, EBI and ENI configuration files for EtherCAT devices, replacing manual creation and cutting setup time by about 80%.
  • Wrote generators for I/O change scripts to automate I/O map updates, reducing manual intervention and error rates when integrating new hardware.
Jun – Aug 2023

Software engineer intern, MicroStep

Bratislava, Slovakia

  • Rebuilt a socket-based application for real-time communication between the CNC control PC and the MMI PC, reducing reliance on internal COBs and lowering latency by about 30%.
  • Remodelled a regex-based file-parsing library used across the CNC automation tools for configuration and log files.
  • Upgraded a Python C-code generator for I/O mapping and device-driver code by integrating that parsing library, cutting manual edits by about 30%.
Jun – Aug 2022

Data analyst intern, MicroStep

Bratislava, Slovakia

  • Analysed and visualised financial and operational data in Power BI to evaluate CNC machine production cash flow, supporting purchasing decisions.
  • Worked with a database specialist to analyse and restructure internal database schemas, improving system efficiency by about 15%.