Speakers
Hendra Rudiansyah & Obrina Briliyant - "Network Packet Security Analysis using SQL-RAG Approach and Cybersecurity-Tuned LLM"
"Keamanan siber modern menghadapi tantangan besar dalam menganalisis data jaringan yang kian masif. Meski machine learning membantu mendeteksi anomali dan mengklasifikasi serangan, analisis mendalam untuk audit maupun forensik masih banyak bertumpu pada pencarian manual dan keahlian manusia. Hambatan ini berpotensi memengaruhi kepatuhan dan bahkan menimbulkan kerugian bisnis.
Presentasi ini menawarkan kerangka kerja baru untuk otomatisasi analisis trafik jaringan dengan memanfaatkan large language model (LLM). Pendekatan ini menggabungkan kueri terstruktur (SQL) dan retrieval-augmented generation (RAG), sehingga pertanyaan sederhana dapat diterjemahkan menjadi wawasan yang cepat, akurat, dan mudah dipahami.
Evaluasi menggunakan RAG assessment suite (RAGAS) diterapkan untuk menilai tingkat akurasi dari output. Hasilnya menunjukkan perbedaan menarik antara LLM umum (base model) dan LLM khusus keamanan siber (specialized model), sekaligus membuka perspektif baru tentang bagaimana AI dapat mendukung organisasi dalam melakukan analisis jaringan secara efisien dan mendalam."
Baskoro Adi Pratomo - "Game untuk Pembelajaran Keamanan Siber"
Under his leadership, ITS Nabu has trained hundreds of participants from academia, government, and industry, and has been recognized by international bodies including the APNIC Foundation and QS Reimagine Education (2024). Baskoro collaborates with partners in the UK, Malaysia, and across Southeast Asia to expand access to cybersecurity education and build a regional cyber defense ecosystem. He is also instrumental in the development of Indonesia’s first cyber range, designed to simulate realistic cyber-attack scenarios for training and evaluation.
"Pengetahuan masyarakat terkait keamanan siber perlu ditingkatkan seiring dengan bertambahnya angka insiden siber, sedangkan persepsi publik terhadap pembelajaran keamanan siber dianggep sebagai sesuatu yang sulit dicapai. Oleh karena itu, kami mengembangkan metode edukasi keamanan siber dalam bentuk aplikasi permainan, yaitu Node Clash (attack-defence), Cryptical: Layer of Protection (kriptografi), dan Malware Adventure (malware). Permainan ini dirancang dengan menyederhanakan konsep-konsep keamanan siber, yaitu malware, kriptografi, dan keamanan jaringan agar mudah dipahami. Pengembangan ketiga permainan dilakukan menggunakan metode Game Development Life Cycle (GDLC). Untuk mengukur efektivitasnya sebagai media pembelajaran, dilakukan eksperimen terhadap dua kelompok: kelompok kontrol yang menggunakan metode pembelajaran pasif (membaca modul) dan kelompok eksperimen yang menggunakan permainan yang dikembangkan. Hasil evaluasi pembelajaran menunjukkan perubahan skor post-test dari pre-test untuk kelompok kontrol banding kelompok eksperimen sebesar 4.4%:11.9% pada Node Clash, 4.57%:23% pada Cryptical: Layer of Protection, dan 29.5%:13.1% pada Malware Adventure."
"Stingrays, also known as IMSI catchers, have captured attention as one of the more mysterious tools in mobile security. By imitating legitimate cell towers, they quietly interact with nearby phones, creating a hidden layer of communication that is rarely visible to the public. Exploring this unseen interaction opens the door to better understanding how mobile networks behave, and how certain signals can reveal more than expected.
The focus of this work is to transform what usually remains invisible—the raw exchanges between a device and the network—into something that can be captured, visualized, and studied. Packet data and signaling messages, once locked away inside complex protocols, become accessible in the form of patterns that highlight the rhythm of mobile communication. With the help of straightforward analysis and a clean dashboard, the signals of the air turn into something tangible and easy to explore.
Instead of presenting dense technical details, the approach highlights the story hidden within mobile signals—what they normally say, how they change, and why those changes matter. By making these hidden interactions more transparent, the research provides not just technical insight but also a new perspective on the invisible world of cellular signals that surrounds us every day."
"DursGo adalah web application security scanner berbasis Go yang dirancang untuk penetration testing dan audit keamanan otomatis. Analisis Berbasis AI: Terintegrasi dengan LLM (Gemini, Groq) untuk memberikan analisis mendalam, ringkasan akar masalah, dan rekomendasi perbaikan kode. Integrasi OAST: Mendeteksi blind vulnerabilities menggunakan verifikasi out-of-band."
"Pembahasan ini bertujuan untuk memperkenalkan Maze: Malware Analysis Platform, sebuah sistem analisis file yang menggabungkan analisis statis, similarity analysis, dan integrasi threat intelligence untuk meningkatkan deteksi serta pemahaman terhadap malware."
Nosa Shandy - "Discovering and Disclosing Privacy Vulnerabilities in Large-Scale Social Media Platforms: Methodology, Impact, and Case Studies"
"This talk presents a practical approach to discovering and responsibly disclosing privacy flaws in major social media platforms. Through three real-world case studies, we show how design gaps can cause data leaks, policy violations, and compliance risks, and discuss their impact on users, trust, and regulation."
In addition to his full-time public service role, Faisal also applies his skills as a freelance penetration tester. This experience allows him to bridge theoretical knowledge with practical, real-world challenges, helping organizations identify and strengthen their digital defenses.
Faisal's professional journey is fueled by a commitment to public service and a belief in the importance of creating a safer digital environment. He is dedicated to contributing to the protection of digital assets and assisting clients, both in the public and private sectors, in navigating the ever-evolving landscape of cybersecurity.
"This research addresses key challenges in securing biometric access devices by proposing a semi-automated security solution that can be integrated into a Telegram bot. We developed and evaluated two scripts, zksentinel-WEB and zksentinel-CLI, to identify vulnerabilities like weak authentication and insecure endpoints in biometric systems. The findings revealed significant risks, including the potential for fraud and unauthorized data access. The results show that this approach is effective in detecting system weaknesses and provides practical guidance for organizations to protect against malicious insider threats and external hackers. This tool is aimed at increasing security awareness and supporting the development of safer technology in the future, helping to minimize material losses and fraud within companies."
