Subject: T. K. Gündüz, MD Speciality: Forensic Science Location: Istanbul, TR Status: Available

Digital forensics & artificial intelligence

Dr. Timur Kaan Gündüz

A forensic medicine physician who spent 17 years at Turkey's Council of Forensic Medicine — twelve as a forensic scientist, then five founding and directing its Forensic Computing Department. I now work independently, applying AI to the parts of investigation that don't scale — the evidence nobody has time to examine.

Practice

Forensics

Digital forensics & incident response

Investigation and analysis where the result has to hold up — under challenge, under scrutiny, in court.

  • Evidence acquisition and integrity
  • Multimedia, CCTV and DVR analysis
  • Data recovery and file carving
  • Expert review of existing findings
  • Methodology, process and lab design
AI

AI systems for investigation

Built, not advised in the abstract. Systems that examine material independently and show their reasoning.

  • Agentic investigation and artifact analysis
  • Automated triage of large evidence sets
  • Multi-model review and adjudication
  • Reporting and case-workflow automation
  • Practical assessment of what AI can and can't be trusted with
Training

Training & enablement

Course material built from real casework, for teams who have to do this themselves.

  • CCTV and DVR forensic video analysis
  • Practical AI for investigative work
  • Lab process and evidence handling
  • Curriculum built for a team's actual caseload

Selected work

Agentic investigation
An agentic DFIR investigation engine — an LLM agent with over 200 tools that parses forensic artifacts, analyses memory dumps, extracts observables, maps findings to MITRE ATT&CK, generates detection rules, and commits every finding to a git-backed evidence chain. Single static binary, no external runtime.
Evidence integrity
A tamper-evident evidence repository — content-addressable storage with cryptographic hashing, a chain-of-custody ledger, versioned findings, deduplication and preserved artifact timestamps. Built so that what you hand a court is provably what you collected.
Multimedia forensics
CCTV and DVR analysis — recovery and examination of proprietary recorder formats, plus an image query engine, and automated review that turns weeks of footage into something an analyst can work through.
Evidence recovery
An advanced file carver — recovering structure and content from damaged, wiped or unallocated media, alongside facial comparison, voice and image analysis systems.
Model adjudication
An adversarial review system — a moderator frames a question as falsifiable hypotheses, a panel of models is split into sides that argue for and against, and a verdict is returned with its reasoning. For questions where one model's answer isn't good enough.
Language & runtime
A semantic graph runtime for LLM-generated programs, executing typed node graphs with provenance traces rather than source files — plus source-to-source compilers targeting Rust from Java, Go and TypeScript.

Record

Certifications & tools

Certifications

  • App Developer for Claris FileMaker Pro — Specialist
  • App Developer for Claris FileMaker Pro — Associate

Working with

  • Cybersecurity
  • JavaScript
  • Go, Rust
  • Godot

Contact

Available for investigations, expert review, training, and building AI systems for forensic work.

Get in touch LinkedIn