Generative AI
Prompt-layer security and containment
Prompt injection, jailbreaks, model misuse, watermarking, alignment, and safe integration remain central challenges.
Structured entry point • Curated topics • Research-depth content
Trustworthy AI systems
A structured portal for navigating AI security across software & hardware.
🔎 Where should I begin? Use the quick links above to jump into domains, foundations, and notes.
AI models
Software security
Hardware trust
Quick links
Start points for different kinds of visitors
These entry blocks make it easy to choose the right next step, whether you want a high-level overview, a direct topic page, a recent note, or a curated reading list.
🧭
Research map
Start with the complete structure of AI security topics and see how the portal is organized.
⚡
Trending topics
Read short, current research-watch notes on emerging papers, attack surfaces, and defense gaps.
📚
Resources
Open curated books, papers, tools, and study paths for structured learning and reference.
👤
About the portal
See the cross-layer perspective, technical strengths, and research background behind this website.
Main research tracks
Three strong entry points into current AI security
These are the three core entry sections highlighted on the homepage. Open the section directly, or enlarge the figure for a quick visual overview before diving in.
Edge AI
Implementation realism at the device boundary
Edge deployment makes physical exposure, firmware trust, side-channel leakage, and constrained-system design highly relevant.
Physical AI
Security, safety, timing, and control together
In embodied systems, sensing, actuation, latency, and cyber risk interact, so trust becomes a full-system design question.
Different research tracks
Browse by security domain
Use these topic links to move directly into the domain that matches your interest, background, or current research question.
Direct topicSoftware securityAdversarial manipulation, poisoning, privacy leakage, extraction, and secure training.
Direct topicHardware securitySide-channel analysis, fault injection, accelerator leakage, and implementation-aware defenses.
Direct topicCloud AI securityShared infrastructure, API exposure, privacy, access control, and secure serving.
Direct topicEdge AI securityPhysical exposure, constrained devices, embedded implementation risks, and deployment realism.
Direct topicPredictive AI securityClassifiers, detectors, ranking systems, forecasting models, and their attack/defense landscape.
Direct topicAgentic AI and countermeasuresAutonomous planning loops, tools, memory, permissions, orchestration, and operational containment.
Cross-layer view
AI security landscape
This visual summary connects model-level vulnerabilities, hardware risks, deployment realities, and countermeasure design into one compact landscape.
Lifecycle perspective
Threats and countermeasures across the AI pipeline
This broader system view complements the landscape above and helps connect software, cloud, edge, hardware, and physical deployment stages into one coherent learning path.
Foundational background
Technical foundations & background
Detailed foundational modules on AI hardware, memory hierarchy, accelerators, inference systems, heterogeneous execution, and distributed infrastructure are organized under the Research section. This keeps the homepage focused while making the deeper technical background easy to access whenever it is needed.
AI Security Pulse
Open pulse page →
Current signals beyond the evergreen research map
This homepage window stays complementary to the Research page by tracking faster-moving developments across academic work, industrial translation, and ecosystem-facing releases.
Academic
Academic Signals
New papers, conference shifts, evaluation trends, and technical ideas that deserve attention before they become part of a static survey.
Full story IndustryIndustry & Innovation
Industrial products, implementation directions, and engineering choices that translate ideas into real systems and deployments.
Full story EcosystemCompanies & Releases
New company moves, platform launches, and product-release signals that matter when AI security leaves the lab and reaches the market.
Full storyBS
About the author
Brojogopal Sapui
A research-focused portal spanning AI security, hardware trust, Edge/Physical AI, with an emphasis on connecting algorithmic concerns to implementation realities.
How to use this portal
Move from overview to depth at your own pace
Start with the research map for structure, open a domain page for focused reading, use the foundations when you want more technical context, and return to Trending Topics for the newest shifts and open problems.
FAQs
Frequently asked questions
These short questions help new visitors understand how to navigate the portal and what kind of material they will find here.
Is this portal for beginners or researchers?
Both. The homepage and topic structure are readable enough for students and collaborators, while the section pages and diagrams remain technical enough for deeper research-oriented reading.
Where should I start if I am new to AI security?
Begin with the research map, then move through software security, hardware security, and edge AI. After that, open the foundations pages whenever you want stronger systems intuition.
What makes this different from a normal personal webpage?
It is structured more like a living learning hub than a simple profile page: topic pages, visual summaries, curated resources, and actively updated research-watch notes all work together.
What should I open if I want the newest material first?
Use Trending Topics. That page is designed to track recent directions and helps complement the more static thematic pages.
Explore more
Explore the portal in the way that suits you
Use the site as an overview page, a structured study map, a technical reference, or a place to follow emerging AI-security directions over time.