Zeeshan Tariq@zeeshantariqpkn

THE PROBLEM CI answers one question: "Did the configured tests pass?" It does not answer the question that matters at merge time: "Is the newly changed behavior adequately verified?" A pull request can add a new branch, a new execution path, or a new edge case while CI stays entirely green — because the existing test suite never exercises the new code. Those tests were written for the old behavior. No test ever fails. The change ships unverified. This is not a test-quality problem. It is a test-coverage blindspot at the point of change, and every team that treats a green CI check as its only safety signal has it. The failure mode is silent and expensive: an untested authorization branch ships on a Friday, and the incident lands on Monday. THE SOLUTION ProofChange is an AI-powered verification layer that closes the gap between "CI passed" and "this change is verified." Its core output is a Change Evidence Package — a structured, machine-readable artifact produced automatically for every pull request. The key insight is ordering. Test generation is step 5 of 7, not the product. ProofChange first proves a gap exists (steps 1–4) and then proves the gap is closed (steps 6–7). The generated test is evidence, not the deliverable. The 8-stage pipeline runs: 1. Diff Analysis — parses the unified diff, extracting changed files, line ranges, and added conditionals. 2. Code Analysis — Python AST walk mapping functions, branches, return paths, and source file paths. 3. Test Mapping — links existing tests to affected functions by name heuristics and docstring scan. 4. Gap Detection — finds branches with no mapped test scenario; promotes security-keyword branches to HIGH severity. 5. Test Generation — IBM Bob produces focused pytest scaffolds targeting each detected gap. 6. Execution — runs real pytest against generated tests, capturing pass/fail/skipped/duration. 7. Impact Analysis — classifies affected paths as verified / partial / unresolved, deduplicated by (source_path, fn_name). 8. Evidence Package — aggregates all outputs into a scored, timestamped JSON + Markdown artifact. TARGET USERS & INTERACTION Backend developers shipping urgent fixes, new developers working in unfamiliar legacy code, and compliance/security teams needing audit evidence. Users interact through a Streamlit dashboard (8 pages) or automatically via a GitHub App webhook that posts a Check Run on every PR — no manual step required. Every PR receives a verdict: STRONG (all gaps addressed, ≥70% paths verified, all tests passed), PARTIAL (some gaps or paths unresolved), or INSUFFICIENT (tests failed or gaps remain). Critically, the level is computed by explicit, auditable rules — not an LLM opinion — and every level carries a plain-language rationale string. WHAT MAKES IT UNIQUE Unlike generic AI test generators that work speculatively across a whole repo from an arbitrary prompt, ProofChange is anchored to the diff itself. It detects gaps before generating, runs tests as a mandatory step, and issues a scored verdict. AI assists the pipeline; the engine runs deterministically without it. IMPACT ProofChange prevents the untested Friday deploy, helps new developers ship safely in unfamiliar code, and gives compliance teams a byproduct artifact — the EvidencePackage JSON is generated on every PR, so audits need no retrospective authoring. Live at proofchange-demo.streamlit.app with a working STRONG-evidence VIP-discount demo, live GitHub PR analysis (App ID 5079526, Check Run posted), 27/27 passing tests, and a green GitHub Actions pipeline.

An AI Decision Intelligence Platform that transforms business data into strategic decisions. Built for CEOs, founders, executives, business analysts, and managers. A complete, production-ready frontend MVP with realistic mock data. Dashboard: 6 KPI cards (Revenue, Profit, Expenses, Customers, Growth Rate, Business Health Score) + 6 charts (Revenue Trend, Sales by Product, Monthly Growth, Regional Sales, Customer Growth, Profit Margin) Data Sources: drag-and-drop upload (CSV/XLSX), dataset preview, delete - Analytics: summary statistics, correlation heatmap, trend analysis, outlier detection, missing values Executive Summary: narrative, SWOT, recommendations, circular Business Health Score gauge

Datapoch Business OS is an AI-powered platform designed to help entrepreneurs, startups, freelancers, agencies, and small businesses transform an idea into a complete business package within minutes. Instead of spending days researching markets, writing business plans, creating marketing strategies, and preparing investor presentations, users simply enter a few business details and let AI generate everything automatically. The platform is powered by Fireworks AI large language models through their OpenAI-compatible API. Fireworks AI is responsible for generating high-quality business plans, landing page copy, marketing strategies, email campaigns, social media content, and investor pitch decks with fast inference and reliable performance. The platform includes an AI Workspace where users can switch between different generated assets without creating multiple prompts. From one business idea, users instantly receive a professional business plan, landing page content, marketing strategy, social media posts, email campaign, and investor pitch deck. The application also includes an analytics dashboard that displays generated businesses, industries, countries, business statistics, and historical activity. Datapoch Business OS was built using a modern technology stack consisting of React, TypeScript, Tailwind CSS, FastAPI, SQLAlchemy, SQLite, Fireworks AI, and OpenAI-compatible APIs. The architecture separates frontend, backend, AI services, database, and workspace modules for scalability. Future improvements include multi-user authentication, cloud database support, collaborative workspaces, AI financial forecasting, competitor analysis, automated website generation, logo creation, CRM integration, and deployment as a multi-tenant SaaS platform for entrepreneurs and agencies worldwide.

The Challenge We Solve: Modern enterprises are data-rich but insight-poor. Business intelligence exists in silos—client data in CRM, financial data in ERP, partner metrics in separate systems. When executives need to make decisions, they're piecing together information from multiple systems with data that's often hours or days old. Our platform solves this by providing a unified, real-time view of the entire enterprise. Our AI-Powered Solution: The AI Executive Intelligence Agent breaks down data silos with five core capabilities: Unified Data Layer: Seamlessly integrates data from all business domains—clients, finance, partners, operations, and competition—into a single real-time view. Cross-Domain Intelligence: Uses AI to identify patterns and correlations across domains. For example: "Client acquisition costs rose because competitor X dropped pricing in Region Y, causing partner referral quality to decline." Predictive Analytics & Scenario Modeling: Forecasts business outcomes and simulates scenarios like "What happens if competitor X acquires competitor Y?" with cross-domain impact analysis. Natural Language Interface: Executives can ask questions in plain English: "Why are acquisition costs rising?" or "Show me partner performance trends." Actionable Recommendations: Goes beyond insights to provide specific, prioritized recommendations with ROI estimates and implementation guidance. Key Features: Real-Time Data Streaming: Processes data as it arrives, not in daily batches Cross-Domain Correlation Detection: Automatically identifies relationships across business functions Predictive Early Warnings: Flags risks before they materialize (e.g., "Q4 target at risk due to competitive pressure") Executive Intelligence Briefing: Daily summaries of key insights, emerging risks, and opportunities Scenario Simulation: Models business scenarios with cross-domain impact assessment Conversational Analytics: Natural language Q&A about business performance

AI Health Guardian is an offline multimodal agentic health scanner designed for low-resource communities where cloud-based healthcare tools cannot operate. It combines vision analysis, cough and breath audio models, rPPG signals, and adaptive questionnaires to produce instant health insights directly on-device—no internet, no servers, no cost barriers. Built for underserved populations, it addresses anemia, respiratory issues, stress, heart-rate anomalies, nutrition risks, and preventive care using edge-optimized ML pipelines. The system also initiates autonomous agentic tasks, such as guiding users, adapting tests, and delivering personalized next-step recommendations. By eliminating cloud dependency and enabling medical screening on any affordable smartphone, AI Health Guardian bridges a critical global gap—bringing privacy-safe, accessible diagnostics to rural, remote, and disconnected environments. This project demonstrates how agentic AI can transform early detection and public health at scale.

GlobalMedTriage is a solution to one of the biggest problems in emergencies: language barriers. When someone calls for help, our system instantly converts their speech into text, translates it into the responder’s language, and provides protocol-driven triage prompts so that first aid instructions can start right away. If the AI is unsure, it uses a confidence check and quickly hands the call to a human interpreter for accuracy. Unlike most telemedicine tools, it also works offline in rural areas with poor connectivity. All calls are encrypted and HIPAA-compliant, ensuring patient safety and privacy. This makes emergency help faster, safer, and accessible to everyone—no matter what language they speak or where they are.

Agentica is a next-generation AI multi-agent orchestration platform designed to revolutionize how businesses and individuals interact with technology. Unlike single-purpose chatbots, Agentica deploys specialized AI agents—each with unique memory, roles, and goals—that collaborate autonomously like a virtual remote team. From fraud detection and phishing protection to customer support, workflow automation, and real-time communication, Agentica transforms complex processes into seamless, intelligent operations. Its modular design allows integration with APIs, AI models, and enterprise tools, making it scalable across industries such as healthcare, finance, and education. With Agentica, organizations gain efficiency, security, and innovation, unlocking the full potential of agentic AI to reshape the future of digital collaboration.

VHeal AI is an Agentic AI-powered hospital discharge copilot designed to streamline patient discharges by autonomously coordinating between hospital staff, systems, and workflows. In modern hospitals, discharging a patient is a multi-step, often delayed process involving doctors, nurses, transport, billing, and pharmacy. VHeal AI detects discharge triggers, reasons through medical workflows, completes checklists, and notifies stakeholders through messaging APIs like WhatsApp and SMS. It also generates discharge summaries using LLMs such as GPT-4o and Claude 3. The project uses a LangChain-based agentic framework, integrated with FastAPI, Streamlit/React, and real-time dashboards for staff. VHeal AI mimics a real hospital staff member, offering timely, autonomous, and safe discharge experiences—helping hospitals save beds, time, and lives.

ReplyMate AI leverages advanced Natural Language Processing (NLP) and machine learning to offer users secure and intelligent conversations. It helps identify potential fraud attempts, ensuring a safer messaging experience on WhatsApp. Built for businesses and individuals, ReplyMate is the go-to tool for fraud prevention and efficient chat automation." Fraud Detection: Identifies suspicious activity and alerts users to potential phishing or scam messages. Smart Replies: Provides contextually relevant and accurate responses to improve user engagement. Real-Time Alerts: Instantly notifies users of fraudulent activities or scams. Security Focused: Ensures that sensitive user data is never compromised during interactions. WhatsApp Integration: Seamlessly integrates with WhatsApp for easy communication and use. Error Handling: Identifies and manages errors gracefully, ensuring the chatbot remains effective even in complex conversations.

Six Matches. One leaderboard. Build on AMD, earn points and compete for prizes.

🌎 Hybrid Hackathon | 💻 Online Build: 3–8 November 2026 | On-site Phase: 8 November 2026 📍 On-site Venue: Pierpont Place 163 W Pierpont Ave, Salt Lake City, UT 84101, USA ✨ Please note that travel and accommodation are not covered.

🌎 Hybrid Hackathon | 💻 Online Build: 12-17 October 2026 | On-site phase: 17-18 October 2026 To participate in the hackathon, simply click Sign up with AMD. If you are not already a member of the AMD AI Developer Program (ADP), you will need to create an account. The AMD Developer Hackathon: ACT III is a hands-on event for developers, founders, engineers and builders who want to push what’s possible with AI on real infrastructure. 📍 If approved for the on-site experience, you'll also have the opportunity to showcase your project live. Travel and accommodation will not be covered.

🌎 Online Hackathon | 💻 Online Build: Sept 26 - Oct 5, 2026 A virtual, global hackathon. Build from anywhere in the world. Dark Factory is the official hackathon of WeAreDevelopers World Congress North America, presented by BAND, but you don't need to be in San Jose to take part. For one week, you build a software factory in BAND Desktop: a band of coding agents that plans work, implements it and checks its own results. 🎟️ Bonus for sign-ups: you are eligible for a free ticket to WeAreDevelopers World Congress North America in San Jose (Sep 23–25). If you do go, stop by the BAND booth: a surprise is waiting for you.

💻 Online Hackathon | ⏳ Build: 48 hours | 📅 September 25–27, 2026 The official IBM Bob 2.0 hackathon. Build what's next in AI-assisted development, alongside the developers and experts shaping it. ⭐ Get help from experts throughout the event 🤝 Join a team or build solo 💰 Prize pool: $12,000 🧑🏻‍💻 Register before the Kick-Off Stream to start building with Bob