sadia usman@Frost_Flashlog3

PromptGuard is an AI security platform built to protect LLM-based applications from adversarial prompts. It analyzes user inputs in real-time and classifies them as Safe, Suspicious, or Dangerous using Meta's llama-3.3-70b-versatile model running on Groq's ultra-fast inference API. What sets it apart is its RAG pipeline — organizations can upload their own security policies, which get chunked, embedded into 768-dimensional vectors using nomic-embed-text, and searched via cosine similarity at query time, so every verdict is grounded in your actual rules. It also includes a policy Q&A chatbot for natural language queries against your knowledge base. Built as a zero-dependency single HTML file and deployed on Netlify, PromptGuard requires no build step and goes live in under 60 seconds.

MoltTrend Claw is a privacy-first autonomous crypto intelligence agent developed for the SURGE × OpenClaw Hackathon. Built on the OpenClaw sovereign, local-first runtime, the agent continuously monitors live cryptocurrency trends using the CoinGecko API and applies advanced AI reasoning through Gemini to detect emerging market narratives. It analyzes sentiment shifts, identifies momentum patterns, and transforms raw market data into structured, actionable insights without requiring constant human supervision. Designed with persistent memory and long-term analytical capabilities, MoltTrend Claw stores historical trend data in JSON to track narrative evolution over time. Through its multi-page Streamlit dashboard, users can explore real-time trend metrics, visual analytics, and historical agent-generated reports. By combining autonomous execution, intelligent forecasting, and structured memory, the project demonstrates how AI agents can support smarter, data-driven crypto decision-making while maintaining user privacy and local control.

SmartStore AI is a simulation-based robotics startup prototype built to address one of retail’s most common and costly challenges: inefficient shelf monitoring and delayed restocking. Retail stores frequently experience out-of-stock situations that lead to lost sales and poor customer experience due to manual and reactive monitoring processes. SmartStore AI solves this by introducing an AI-driven system that continuously monitors shelf status, tracks customer traffic intensity, and measures how long shelves remain empty. Using a priority-based decision engine, the system automatically generates restocking tasks and simulates an autonomous robot executing them through a structured workflow (Dock → Shelf → Restock → Return). Built with Python and Streamlit, the platform includes a live interactive dashboard that displays shelf conditions, traffic levels, robot position, restocking activity, and performance metrics in real time. Designed as a simulation-first solution, SmartStore AI validates autonomous retail operations in a safe and scalable environment, while laying the foundation for future expansion into multi-robot systems, digital twin environments, computer vision integration, and demand forecasting. The project represents a production-minded, startup-ready robotics solution that bridges AI, automation, and retail intelligence.

DevBug-AI is an AI-powered system that automates bug classification and developer recommendation for software teams. As applications scale, bug triage becomes slow and error-prone due to unclear reports, inconsistent categorization, and manual assignment. DevBug-AI removes this bottleneck by intelligently analyzing bug reports and assigning them to the most suitable developers. Using natural language processing and machine learning, the system classifies bugs from their title and description and recommends the top three developers based on historical bug data, domain, and tech stack. Confidence scores are provided to support reliable decision-making. The solution is built as a Streamlit web app with a unified ML pipeline. It uses TF-IDF and sentence embeddings for text understanding, Scikit-learn for bug classification, and LightGBM for developer recommendation. The system is trained on 50,000+ real bug reports and gracefully handles unseen technologies by mapping them to an “Other” category. DevBug-AI reduces triage time, improves assignment accuracy, and enables data-driven bug management at scale. Future work includes integrations with Jira and GitHub, workload-aware recommendations, and LLM-based bug summarization—making it a practical, production-oriented AI solution for modern engineering teams.

AgentPay-AI is a proof-of-concept platform that demonstrates how Generative AI services can be monetized using pay-per-use, token-based billing—similar to real-world AI APIs. Built with Streamlit and Google Gemini, the system simulates a USDC-style wallet that estimates token usage, deducts balance per request, and only executes AI tasks when sufficient funds are available. This project addresses a major gap in AI demos: cost transparency and usage accountability. AgentPay-AI showcases how AI-as-a-Service (AIaaS), agent marketplaces, and crypto-enabled AI platforms can implement realistic billing logic. Key Highlights: Token-based cost estimation Simulated USDC wallet per session Controlled AI execution based on balance Google Gemini / PaLM integration Simple, intuitive UI Designed as a hackathon and portfolio project for GenAI, SaaS, and Web3 applications.

An AI-powered interactive web application built with Streamlit that predicts whether a candidate will get placed in a job (or admitted) based on academic performance and other features. The model simplifies decision-making for students, HR teams, and academic advisors by providing data-driven placement predictions. The user inputs academic and background features, including: SSC percentage HSC percentage Degree percentage MBA percentage Work experience Specialization Gender And more Inputs are one-hot encoded for categorical features. A Logistic Regression model (trained offline) is loaded using Pickle. The model outputs a binary prediction: “Placed” or “Not Placed”. The result is displayed on the Streamlit app in a clear, user-friendly format.

AskTheWeb: AI-Powered Website Question Answering System AskTheWeb (also known as WebMind AI) is an advanced Question-Answering application designed to tackle information overload by transforming static websites into interactive, conversational experiences. Built using Streamlit, the application leverages the speed and intelligence of Google Gemini 2.0 Flash to understand and synthesize web content in real-time. The core functionality relies on a robust RAG (Retrieval-Augmented Generation) pipeline designed for accuracy and persistence. When a user inputs a URL, the system employs Requests and BeautifulSoup to scrape and clean the HTML data, acting as an efficient ETL transformer to remove messy code and script tags. This cleaned text is converted into high-dimensional vector embeddings using Google’s GenAI SDK and stored in ChromaDB, which acts as the application's "Long-Term Memory". This architecture allows users—such as students, researchers, and analysts—to ask complex natural language questions and receive instant answers that are grounded specifically in the context of the provided URL. Unlike standard chatbots, AskTheWeb ensures that answers are relevant to the specific source material provided. We also addressed significant technical challenges during development, specifically ensuring compatibility with cloud environments. We implemented a custom solution using pysqlite3-binary to patch SQLite version incompatibilities on Streamlit Cloud, ensuring the vector database runs smoothly in production. The result is a scalable, modular tool that makes researching the web faster and more intuitive.

AgentFlow: AI for Productivity A smart to-do list that prioritizes tasks into 4 quadrants: Urgent & Important | Important, Not Urgent | Urgent, Not Important | Not Urgent, Not Important AI agents provide instant resources & tools for each task. Users can add custom agents or discover new ones from lablab.ai How It Works Tech Stack: Frontend: Next.js + TailwindCSS AI: Gemini 2.5 Flash Deployment: Vercel Process: User creates a task. AgentFlow assigns it to the right priority box. AI recommends best agents/resources (e.g., blog writer, code assistant). Task can be tracked, updated, or deleted once done.

🩺 Problem Statement Medical students often struggle to bridge the gap between theory and practice. While textbooks provide knowledge, students rarely get the chance to practice clinical reasoning, patient interviews, and decision-making in a safe, low-stakes environment. Traditional case-based learning is often static, repetitive, and lacks interactive feedback, leaving students unprepared for real-world patient encounters. 💡 Our Solution: DocQuest DocQuest is an interactive AI-powered simulation platform where learners can: Interview virtual patients powered by GPT-5. Practice diagnostic reasoning by proposing tests, forming differential diagnoses, and planning management. Receive instant AI-driven feedback with scores, learning points, and red-flag highlights. It’s like a safe “practice ward” where students can make mistakes, learn, and improve — without any risk to real patients. 🚀 Key Features Case Library: Browse cases across multiple specialties (Cardiology, Neurology, Respiratory, Endocrinology, etc.). AI Patient Interview: Ask questions and interact dynamically with a simulated patient. Solve Cases: Enter diagnosis, key tests, and management plans. Feedback & Scoring: AI evaluator provides structured feedback (Diagnosis 4/4, Tests 3/3, Plan 3/3). Progress Tracking: Track cases completed, average score, and improvement over time. Today’s Challenge: Daily highlighted case for gamified learning. 🔮 Future Additions Leaderboard & Badges: Gamify the experience with rewards and peer competition. Multiplayer Mode: Allow group discussions on a case with collaborative solving. Custom Case Builder: Let educators create and upload their own cases. API Integration: Connect with medical knowledge APIs (e.g., UpToDate, PubMed) for deeper references. 👉 Our vision is to empower medical students and young doctors worldwide with a safe, engaging, and effective tool to sharpen their diagnostic reasoning skills.

📅 June 12–19, 2026 • June 12: Kick-off Stream • June 12–19: Online Build Phase • June 19: Project Submissions Ends 🌍 Fully online hackathon. Join and build from anywhere in the world. 🤝 Build practical enterprise workflows where AI agents communicate, coordinate, exchange context, and complete work together. 🧠 Powered by Band and Codeband, built for collaborative AI systems and agent orchestration. 🏆 Prize Pool: TBA 🧑‍💻 Apply now to build the next generation of enterprise multi-agent systems.

📅 May 11–19, 2026 • May 11 – 19: Online Build Phase. • May 18 (Hybrid Build Day – Online & Onsite) • May 19 (Demos & Awards – Online & Onsite) 🌍 Hybrid event. Join online from anywhere or build onsite at the venue. 🏆 Winners will be announced live on stage at the AI & Big Data Expo North America on May 19 📍 On-site Venue (May 18–19): San Jose McEnery Convention Center, CA, USA 🤝 Join solo or with a team. 🎟 FREE AI & Big Data Expo North America ticket included for participants that are able to join the on-site phase. 🏆Compete for a $10,000 prize pool 🧑‍💻 Apply now to design, prototype, and showcase enterprise-ready innovation.

Grab your ideas, spin up compute, and start building what’s next! Start with $100 in AMD Developer Cloud credits and begin building right away. Get access to hands-on tutorials, workshops, and curated learning resources to help you build faster and go further. To unlock credits and join the hackathon, sign up for the AMD AI Developer Program. 🤝 Go solo or team up — ship something real. 💰 $21,500+ total prize pool & 1 × AMD Radeon™ AI PRO R9700 GPU up for grabs. 📍 On-site Venue (May 9-10): MindsDB SF AI Collective 3154 17th St, San Francisco, California, USA 📍On-site participation is by invitation only. Travel and accommodation expenses will not be covered. 🧑‍💻 Developers, builders, and innovators - bring your boldest ideas and let’s turn them into real AI systems on AMD.

Join our $1,000,000+ startup challenge series powered by Surge 📌 This announcement outlines the launch of a Global Human+AI Exchange — aligning universities, global cities, talent networks, and industry to accelerate human-centered AI innovation, technology transfer, and economic growth. 👉 Read more here ⏱️ 8 days to turn your idea, or existing product, into an investable demo. 📅 February 6 - 15, 2026 • Feb 6 - 14 (Online Phase) - Collaborate and build online with developers and AI innovators from around the world. All projects must be submitted by the end of the online phase on February 14th. 🕙 Doors open at 10:00 AM, first come, first served. • Feb 14 (On‑site Build Day) - Selected participants will be invited to an exclusive in‑person session to refine their projects and connect directly with mentors. • Feb 15 (On‑site Demos & Awards) - Live pitching sessions to a panel of judges and ecosystem partners, followed by the official winner announcement. 🌟 Get feedback from startup mentors and technical experts. 📍 On-site Venue (Feb 14–15): MindsDB SF AI Collective 3154 17th St, San Francisco, California, USA 📲 Real-time on-site updates (SF) For real-time announcements and information during the on-site portion (Feb 14–15), join the LabLab SF Chapter WhatsApp group: 👉 Join the WhatsApp group 🤝 Join solo or with a team. New founders and existing startups are welcome. 🏆 $1,000,000+ in credits, token prizes, perks and funding opportunity. 📍 On-site participation is by invitation only. Travel and accommodation expenses will not be covered. 🧑‍💻 Apply now to build, validate, and pitch with purpose.