GLOG.AI® Scientific Innovation • Deep-Tech Research & Development

Pioneering the Future of AI-Driven Software Security, Predictive Threat Intelligence, and Autonomous Cyber Defense

Explore Glog.AI’s research and development initiatives, encompassing cybersecurity, software security, artificial intelligence, and broader scientific and technological advancements.

Our strength lies in the collective intelligence and passion of our team. Comprised of a combination of seasoned and young cybersecurity experts, data scientists, AI/ML engineers, and security analysts, we share a common vision: to create a safer digital world. We foster a culture of collaboration, continuous learning, and relentless pursuit of excellence. Each member brings a unique perspective and a wealth of experience, allowing us to tackle the most challenging security problems with innovative solutions. We are an international team with the right balance of industry professionals from cybersecurity, innovators, and scientists who possess deep knowledge in cybersecurity—specifically in software and application security as well as machine learning and AI.

100+ Papers & 10 Books Foundational peer-reviewed research in AI, cryptography, network & software security
#1 on RealVuln v3.1 82.7 F1 Score | 85.1% Precision | 80.5% Recall via whole-application semantic graphs
10.5 Gbit/s AI Sensor Passive wire-speed network behavioral modeling on 2 vCPU with zero payload export
Global Standards Active contributions to IEEE AI Standards Working Groups & recommended in the Geneva Manual

Core GLOG.AI® Research & Development Pillars

How our scientific breakthroughs translate directly into the unified Glogosphere® cybersecurity ecosystem.

1. Semantic Graphs & Auto-Remediation

Combining deterministic whole-application data/control-flow graph construction with targeted AI reasoning to eliminate false positives and automatically generate context-aware code patches.

2. Wire-Speed Behavioral Network AI

Ultra-lightweight, out-of-band stream processing and machine learning models (NetEcho™) that classify 10+ Gbit/s traffic and 650k SYN flood flows/sec without touching packet payloads.

3. Predictive Threat Modeling (SP)

Probabilistic forecasting models that correlate global exploit telemetry, CVE evolution, and asset exposure into a single business-weighted Asset Risk Score (0–100).

4. Trustworthy AI & Global Standards

Advancing international standards for AI risk, safety, LLM evaluation, and agentic security across IEEE working groups, ACM SIGSAC, ISACA, and academic institutions.

Scientific Recognition, Publications, Papers & Conference Presentations

Click on any section below to explore our published research, academic citations, and conference papers.

1. Scientific Leadership & Academic Profiles — Dragan Pleskonjić

Three Decades of Scientific Research, Patents, Textbooks, and Deep-Tech Entrepreneurship

Glog.AI is initiated, founded, and led by CEO Dragan Pleskonjić. Visit his personal website (dragan-pleskonjic.com) for more details about his broader scientific research and pages: About Me, Resume, Patents, Books, Events, Publications, Projects, and more. Additionally, his website features an extensive archive of events, keynote presentations, and videos where he spoke and presented.

Explore Academic & Research Profiles

2. International Monograph Recognition — “Transformation of the insurance market: responses to new challenges”

Glog Highlighted as a Leading-Edge AI Software Cybersecurity Provider in Cyber Insurance

The monograph, “Transformation of the insurance market: responses to new challenges,” an internationally significant scientific publication, mentions Glog under its “RISK AND COMPLIANCE – Cyber Insurance” section. It highlights Glog as a company that provides AI-driven, leading-edge solutions for software cybersecurity.

Excerpt — Section 6.2. Cyber Insurance

“With the increasing prevalence of cyber threats and data breaches, cyber insurance has become a critical component of risk management for insurers and policyholders alike. AI enhances cyber insurance offerings by enabling insurers to assess cyber risks more accurately and comprehensively. By analyzing cyber threat intelligence, historical data, and industry trends, AI algorithms can identify emerging threats, quantify potential losses, and tailor cyber insurance policies to the specific needs and risk profiles of individual clients. In Serbia, for example, the company called Glog is providing AI-driven leading-edge solutions for cyber security of software.438 The proactive approach to risk mitigation not only helps mitigate financial losses associated with cyber incidents but also fosters resilience and confidence in the face of evolving cyber threats.”

— Footnote 438: https://www.glog.ai/

Full Bibliographic Source

Jelena Kočović, Biljana Jovanović Gavrilović, Tatjana Rakonjac-Antić, Branislav Boričić, Zorica Mladenović, Marija Koprivica, Transformation of the insurance market: responses to new challenges: [monograph of international significance]. Belgrade: University of Belgrade, Faculty of Economics and Business, Publishing Centre, 2024. [Online]. Available: https://cid.ekof.bg.ac.rs/wp-content/uploads/2021/02/Transformation-of-the-insurance-market.pdf

3. Global Cyber Governance & Book Chapters — The Geneva Manual, “Serbia AI Landscape 2026” & “Digitalna prizma”

International Recognition for AI-Driven Vulnerability Auto-Remediation and Cyber Resilience

  • Recommended in the International Geneva Manual: GLOG.AI® is directly cited and recommended in the Geneva Manual for responsible behavior in cyberspace, highlighting its AI-driven methodology for open-source software vulnerability detection and automated remediation. (Read article →)
  • Named a Top 10 AI Innovator (2026): Recognized among the Top 10 AI Innovators in the “Serbia AI Landscape 2026” report by the Global Investors Forum.
  • Multidisciplinary Book “Digitalna prizma”: Authored the dedicated “Cybersecurity” chapter examining the intersection of artificial intelligence, software security, and digital transformation.
  • IEEE Standards Working Groups: Active participation in global standardization efforts including IEEE P3396™ (AI Risk, Safety, Trustworthiness, and Responsibility), IEEE P3394™ (LLM Agent Interface), and IEEE P3419™ (LLM Evaluation).
4. Featured Research Paper — “AI’s Role in Cybersecurity Threats and Defenses – Examples of Concrete Solutions” (INFOTECH)

INFOTECH – Government Technologies • International Conference & Exhibition

  • Paper Title: “AI’s Role in Cybersecurity Threats and Defenses – Examples of Concrete Solutions”
  • Authors:
    • Dragan Pleskonjić, Glog.AI
    • Luka Tica, Glog.AI
    • Vladimir Jelić, Glog.AI
    • Dušan Todorović, Glog.AI
    • Anthony English, Bot Security Solutions Inc.
  • Conference & Venue: INFOTECH 2025, Hotel Izvor, Aranđelovac, Serbia (June 4, 2025)
  • Publication Details: INFOTECH 2025 Proceedings — ISBN 978-86-900491-3-4
Interested in the full paper or our applied AI defense methodologies? Contact our Research Team →
5. 2025–2026 Keynotes, Conference Presentations & Executive Media Broadcasts

2026 Featured Keynotes, Panels, Live Demonstrations & Broadcasts

Katapult Demo Day & Silicon Valley Showcase — San Francisco, USA (November 2026)

  • Date & Location: November 4, 2026 • San Francisco & Palo Alto, California, USA
  • Overview: Executive presentation and technical showcase of the unified GLOG.AI® Glogosphere ecosystem—demonstrating #1-ranked automated source code vulnerability remediation (RealVuln v3.1) and passive 10.5 Gbit/s AI network anomaly detection (NetEcho™) to North American enterprise leaders and technology investors.
  • Speaker: Dragan Pleskonjić, Founder & CEO, GLOG.AI®

Bloomberg Adria TV — “AI Thinkers” Expert Broadcast Series (August & September 2026)

  • Program & Host: Bloomberg Adria TV’s flagship series “AI Thinkers” hosted by Stefan Dević
  • Episode 1 (August 2026) — “AI in Cybersecurity: Shield or Weapon?”: In-depth analysis of how offensive AI and autonomous agents accelerate cyber attacks—and how defensive neuro-symbolic AI, semantic code graphs, and predictive telemetry tip the balance back to defenders.
  • Episode 2 (September 2026) — “The Cost of AI Control”: Examining the economics and governance of enterprise AI adoption, preventing Shadow AI and data leakage, enforcing digital sovereignty, and aligning with the EU AI Act and international safety standards.

Machine Learning Conference (MLC2026) — Subotica, Serbia (May 2026)

  • Date & Organizer: May 2026 • Subotica, Serbia (Hosted by InspiraHub)
  • Keynote & Live Demonstration: Delivered a technical keynote presentation and live software security demonstration showing how GLOG.AI® constructs whole-application semantic graphs to eliminate false positives and automatically synthesize context-aware remediation patches in real time.
  • Speaker: Dragan Pleskonjić, Founder & CEO, GLOG.AI®

Čikom Cyber Security Days 2026 — Podgorica, Montenegro (May 2026)

  • Date & Venue: May 2026 • Hotel Hilton, Podgorica, Montenegro
  • Expert Panel & Presentation: Featured panelist addressing proactive AI cyber defense, software supply-chain assurance, and high-speed passive network anomaly detection for financial, telecom, and public-sector infrastructure.

Pupin Connectivity and Investment Forum 2026 — Belgrade, Serbia (May 2026)

  • Date & Venue: May 2026 • Hotel Hilton, Belgrade, Serbia
  • Focus: Trans-Atlantic deep-tech collaboration, sovereign AI infrastructure, and enterprise cybersecurity innovation.

2025 Featured International Conferences, Case Studies & Interviews

Black Alps Cybersecurity Conference 2025 — Yverdon-les-Bains, Switzerland (November 2025)

  • Date & Location: November 2025 • Yverdon-les-Bains, Switzerland
  • Focus: International cybersecurity research exchange on applied artificial intelligence, automated vulnerability remediation, and sovereign European cyber defense.

11th International Forum “Safety for the Future” 2025 — Belgrade, Serbia (September 2025)

  • Date & Venue: September 2025 • Hotel Srbija, Belgrade, Serbia
  • Focus: AI-driven critical infrastructure protection, predictive threat intelligence, and national cyber resilience.

Be Risk Protected Conference 2025 — “Risks of the New Era and Sustainability – Flight into the Unknown”

  • Date & Venue: June 5, 2025 • Hyatt Regency, Belgrade
  • Case Study Presentation: KAKO DA UZ POMOĆ AI “UBIJETE AŽDAJU” — Konkretni alati: kako da se najefikasnije odbranite od modernih AI sajber napada (How to “Slay the Dragon” with AI — Concrete Tools: How to Most Effectively Defend Against Modern AI Cyber Attacks)
  • Speaker: Dragan Pleskonjić, Founder & CEO, GLOG.AI®

Cybersecurity Summit 2025 — Proactive Threat Intelligence & Autonomous Agents

  • Featured Session: From Reactive Defense to Proactive Security: The Future of Threat Intelligence — Explores how threat probability models, behavioral analytics, and deep network traffic analysis—combined with AI and integrated security platforms—empower SOC teams to detect, prioritize, and neutralize threats before they cause harm.
  • Use Cases & Challenges: Practical examples of integrating security platforms and overcoming unified threat intelligence complexities.
  • Published Interviews:
    • Dragan Pleskonjić on AI, Cybersecurity, and the Age of Autonomous Agents – Interview with Cybersecurity Summit (Part 1, April 14, 2025)
    • Dragan Pleskonjić on AI, Cybersecurity, and the Age of Autonomous Agents – Interview with Cybersecurity Summit (Part 2, April 17, 2025)
See more events, videos, and conference updates on our Events & News and Resources pages →
6. Empirical R&D Validation — RealVuln v3.1 (#1 F1 Leaderboard) & NetEcho™ 10+ Gbit/s Benchmark

Where Scientific Theory Meets Verifiable, Reproducible Benchmark Performance

  • #1 on the RealVuln v3.1 Leaderboard (September 2026): Across 34 evaluated scanners and 66 open-source Python applications (133,782 lines of code; 1,902 labelled vulnerabilities), GLOG.AI® Software Security ranked #1 with an 82.7 F1 score (4.4 points ahead of the next scanner), achieving 85.1% precision and 80.5% recall—the only scanner to clear 80% on both simultaneously while producing 47% fewer false positives. (Read full benchmark report →)
  • GLOG.AI® NetEcho™ 10+ Gbit/s Wire-Speed Benchmark (October 2026): Verified 93.9%–100% passive packet capture up to 10.5 Gbit/s (~890,000 packets/sec) and resilience under 650,000 SYN flood flows/sec on a single 2 vCPU / 2 GiB RAM container across 3,322,311 sessions and 1,727,099 anomalies—with zero dropped records, 7.5s average AI detection latency, and zero packet payloads leaving the network. (Read full NetEcho™ benchmark →)

From Scientific Research to Commercial Impact

How GLOG.AI® R&D initiatives power our commercial products and enterprise services.

Research Domain Core Scientific Innovation GLOG.AI® Product / Service Verified Outcome & Recognition
Program Analysis & Neuro-Symbolic AI Whole-application semantic graphs (data & control flow) paired with contextual AI exploitability judgment and patch synthesis GLOG.AI® Software Security #1 on RealVuln v3.1 (82.7 F1); −47% false positives; proven on 343 AMD ROCm™ repos & Easy Aerial DoD Blue UAS clearance
High-Speed Stream Processing & Behavioral ML Zero-copy out-of-band flow aggregation and unsupervised/supervised anomaly classification without payload inspection GLOG.AI® NetEcho™ 10.5 Gbit/s capture on 2 vCPU; 650k SYN flood flows/s resilience; 7.5s average latency; 100% sovereign telemetry
Probabilistic Threat Modeling Multi-source threat intelligence fusion and predictive weaponization forecasting Security Predictions (SP) & vSOC Unified 0–100 Asset Risk Score correlating code vulnerabilities, threat probability, and live network traffic
AI Safety, Governance & Agentic Security Defense against agent goal hijacking, tool misuse, memory poisoning, hallucinations, and data leakage GLOG.AI® Trainings & Advisory Aligned with IEEE P3396/P3394/P3419, EU AI Act, OWASP Agentic Top 10, and the Geneva Manual

Partner with GLOG.AI® on Applied Research or Enterprise Deployment

Whether you are looking to deploy our #1-ranked AI cybersecurity solutions in your enterprise, collaborate on European and international research consortia (Horizon Europe, EIC, Digital Europe), or invite our leadership to speak at your conference—we would love to hear from you.