Criminal Procedure • Audio Forensics • Digital Evidence • Data Analysis
I use GitHub as a learning laboratory and a platform for transparent, reproducible technical-legal research.

I work across the intersection of:

• criminal procedure and evidence law,
• audio forensics (diarization, transcription, signal analysis),
• digital evidence engineering and chain-of-custody reconstruction,
• technical building assessments (heating systems, thermal audits, documentation).

My repositories document how law, technology, measurement, and methodical reasoning can be integrated into reproducible workflows.

GitHub allows me to:

• experiment with tools such as Python, Whisper, pyannote, ffmpeg,
• build full evidence repositories for real cases (e.g., II K 70/24),
• teach others how technical and legal inquiry complement one another.

Within GitHub Education and the Student Developer Pack, I focus on:

• building open case studies for students of law, forensics, and data science,
• creating step-by-step pipelines for evidence analysis:
  • audio preprocessing,
  • diarization and transcription,
  • timing consistency checks,
  • comparison between courtroom recordings and official transcripts,
• documenting research-ready, reproducible workflows.

My repositories are designed so that anyone can:

1. clone the project,
2. follow the README instructions,
3. reproduce the analysis with full transparency.

• advanced diarization and speaker-tracking (pyannote),
• forensic-grade signal analysis (formants, silence segmentation, artifact detection),
• automated procedural reporting (Python → DOCX/PDF pipelines),
• GitHub Actions for:
  • auto-updating statistics,
  • generating forensic summaries,
  • validating data integrity in multi-file repositories.

A structured, multi-layered reproduction of a real criminal case, including:

• evidence indexing (“digital case file” format),
• audio recordings (M4A/OGG/WAV) + diarization,
• transcript comparison tools,
• chain-of-custody reconstruction,
• educational notebooks explaining the methodology.

Intended use: legal tech courses, forensic audio workshops, digital evidence methodology training.

A modular pipeline for:

• audio conversion and resampling (ffmpeg),
• diarization (pyannote),
• Whisper-based transcription,
• formatting transcripts for legal review.

Includes well-commented workflows intended for beginners and advanced students.

A technical repository documenting:

• heating & hot-water system diagnostics,
• 3D measurement workflows (Leica DISTO X6),
• thermal imaging interpretation (FLIR C5),
• building audit methodology.

Intended use: building-science labs, civil engineering students, interdisciplinary research combining law & engineering.

Languages & Tools

• Python (analysis, automation, reporting)
• ffmpeg, sox (audio processing)
• Whisper, pyannote-audio (speech & diarization)
• Git, GitHub (Actions, LFS, structured evidence repositories)
• Audacity / Adobe Audition (signal inspection)

Hardware

• Dell Precision 7550 (CUDA)
• Samsung Galaxy S24 Ultra (raw recordings, photogrammetry)
• Leica DISTO X6, FLIR C5, Bosch GLL 3-80 C
• HP Color LaserJet Pro M252dw (documentation output)

This profile uses GitHub Actions to automatically generate:

• commit-activity graphs,
• language-usage statistics,
• repository analytics.

Example outputs:

The following exercises are designed for students, educators, and researchers who want to explore digital forensics, legal-tech analysis, or data-driven methodology.

Objective: Reproduce a basic forensic workflow.
Tasks:

1. Clone fonoskopia-tools.
2. Convert the sample audio files using ffmpeg.
3. Run diarization and generate a segment map.
4. Compare diarization timestamps with the transcript.
5. Submit a short report explaining discrepancies.

Objective: Identify procedural anomalies.
Tasks:

1. Load a provided courtroom audio segment.
2. Generate a Whisper transcription.
3. Compare it line-by-line with the “official transcript”.
4. Highlight mismatches related to:
   • omission,
   • misattribution of speakers,
   • altered sequencing.
5. Discuss the impact on fair-trial guarantees.

Objective: Understand digital evidence lifecycle.
Tasks:

1. Navigate the IIK_70_24 directory structure.
2. Review metadata in the metadane/ folder.
3. Build a timeline of evidence creation, copying, and storage.
4. Identify any “breaks” in the chain.
5. Submit a formal structured report.

Objective: Apply engineering measurement workflow.
Tasks:

1. Review photographic and thermographic documentation in joliot-curie-19a.
2. Analyze heat-loss indicators and insulation patterns.
3. Correlate sensor readings with structural documentation.
4. Produce a short engineering-legal assessment.

Objective: Learn reproducible research using CI.
Tasks:

1. Copy the provided workflow.
2. Configure a daily stats updater.
3. Add a Python script generating a PDF summary.
4. Publish results to the repository’s README.

I welcome collaboration from:

• students and researchers in law, digital forensics, signal processing,
• engineering and data-science programs,
• instructors looking for real-world, reproducible case studies.

For academic or research inquiries:
michal@trojaczek.com

“Theory ends where evidence begins. The rest is documentation.”