Timur Biyachuev
Independent Researcher, AI Agent Security
Moscow | timur@biyachuev.com
Current Research
- Security and threat modeling of AI agents and agentic systems
- Prompt injection, indirect prompt injection, and untrusted-content handling
- Agent identity, permissions, tool access, and trust boundaries
- Security of RAG corpora, persistent memory, and downstream AI readers
- Runtime controls, observability, incident investigation, and agent lifecycle governance
- Ambient and embodied AI security
Research and Academic Background
Independent Researcher, AI Agent Security
2026 – present
Independent research practice on the security of agentic AI systems.
MIPT Business School – Academic Advisor
2026 – present
Advisor to the Master’s program in Technology Entrepreneurship. Mentoring projects in computer vision, AI assistants, and product strategy.
Huawei Russian Research Institute – Director, Moscow Basic Software Lab
2022–2023
Took over a team of approximately 60 engineers and grew it to 120 across Moscow, Novosibirsk, and St. Petersburg as part of the Central Software Institute: operating systems, compilers, AI frameworks. Established partnerships with academic institutions.
Samsung Research Russia – Head of System Software Development
2021–2022
60 engineers across eight research and engineering areas, including Android Runtime, TEE operating systems, Rust with ARM CCA, and LLM-based machine learning on code. Opened two new research directions and secured headquarters funding for them through internal grant competitions.
Kaspersky – Vice President, Threat Research
2018–2019
Threat research organisation of 300+ people, including 110+ researchers in the anti-malware and anti-spam laboratories. Established the company’s AI/ML competence centre: reputation-based detection, threat classification, and on-device behavioural models for endpoints.
Kaspersky – Director, Anti-Malware Research Lab
2014–2018
Led the laboratory responsible for continuous protection of 100M+ users. Built the automated threat detection pipeline – sample collection, sandbox and similarity analysis, machine learning classification, detection generation. The share of threats detected automatically grew to 99.9%.
Kaspersky – Head of Infrastructure Development / Head of Technology Projects, R&D
2006–2014
Automation systems of the anti-malware laboratory: sample processing, classification, and the machine learning based detection technologies that grew out of it.
ITMO University – Department of Secure Information Technologies
2001–2006
Progressed from laboratory assistant to associate professor. Candidate of Sciences (Ph.D.) in Information Security, 2005. M.Sc. in Computer Engineering, honours degree, 2003. Established a joint research laboratory with Kaspersky Lab.
Patents
Automatic description and clustering of malicious objects. RU 91213 U1 (registered 2010) and US 8,635,694 B2 (granted 2014) – one invention in two jurisdictions, shared priority RU 2009136235. Co-inventors: A. Malyshev, T. Biyachuev, D. Ilin. Assignee: Kaspersky Lab.
Anti-phishing system. RU 103643 U1 – filed 01.11.2010, published 20.04.2011. Co-inventors: A. Malyshev, T. Biyachuev.
Full career history, including business and product leadership roles, is on the About page.