kanish = {
"name" : "Kanish Tyagi",
"location" : "Arlington, TX 🇺🇸",
"education" : "M.S. Data Science @ UT Arlington (Dec 2026) — GPA 3.8",
"current_role" : "AI Business Analyst Intern @ Triple-S Steel Supply LLC",
"background" : "3+ yrs QA Engineer → Data Scientist → AI Agent Builder",
"building" : "AI Copilot Orchestrator (Copilot Studio + Azure AI Foundry)",
"open_to" : ["AI Product Manager", "Data Scientist", "AI Business Analyst"],
"superpower" : "Turning failure-mode thinking into production AI systems 🧠",
}- 🔭 Working on: Multi-domain AI Copilot Orchestrator routing NL queries across 5 Fabric AI Skill agents — cutting response time from 5-6 min sequential to 48 seconds
- 👯 Collaborating on: AI agent workflows, LLM-powered automation, Power Platform solutions, and agentic system design
- 🤝 Looking for: AI PM, Data Scientist, or AI Business Analyst roles — open to referrals and networking!
- 🌱 Learning: On-behalf-of (OBO) auth flows, Azure AI Foundry advanced orchestration, and Power BI RLS architecture
- 💬 Ask me about: AI agent design, Copilot Studio, Power Automate, Python, XGBoost, or my QA → AI journey
- ⚡ Fun fact: Cut post-release defects 30% as a QA Engineer. Now I build AI agents that process 3,000+ rows in 15 minutes — unattended. Debugging prod AI > debugging prod code 🚀
| 🎤 Speaker | 🏆 Hackathon | 🔀 Open Source | ⚙️ Production AI |
|---|---|---|---|
| UTA Business Analytics Symposium 2026 | 2nd Runner-Up — AWS GenAI Hackathon | 7 PRs Merged @ Microsoft | Nesting Bot: 98-99% accuracy, 3,000+ rows/run |
| Project | Stack | Highlights |
|---|---|---|
| 🤖 AI Copilot Orchestrator (Triple-S Steel) | Copilot Studio · Power Automate · Fabric AI Skills · SharePoint | 5 parallel agents · 48s response (from 5-6 min) · Config-driven |
| 🏗️ Nesting Bot (Triple-S Steel) | Copilot Studio · Azure AI Foundry GPT-4o · Office Scripts · Power Automate | 5 input formats · 98-99% accuracy · Live production daily |
| 📬 Invex Address Automation (Triple-S Steel) | Power Automate Desktop · Excel · SmartClient/GWT | 626 records · 7 hrs unattended vs 3 people/2 weeks |
| 📚 Adaptive AI Study Tutor | Python · UCB1 RL · Claude API · Streamlit · SQLite | 18 config arms · Real-time adaptation · Full-stack |
| 🏥 Health Center Attrition Predictor | XGBoost · Scikit-learn · Pandas · GridSearchCV | 87% accuracy · 1,400+ records · 3 models compared |
| 📊 GenAI ROI Analysis | Python · Pandas · Matplotlib · Survey Analytics | 60,907 responses analyzed · Quantified GenAI adoption ROI |
| 👥 Customer Segmentation | K-Means · PCA · Scikit-learn · Seaborn | 4 segments · Behavioral clustering · Marketing insights |
| Badge | Certification | Issuer | Year |
|---|---|---|---|
| ☁️ | Azure AI Fundamentals (AI-900) | Microsoft | 2026 |
| ⚡ | Power Platform Fundamentals (PL-900) | Microsoft | 2026 |
| 📊 | Data Analytics Professional Certificate | 2025 | |
| 📋 | Project Management Professional Certificate | 2025 | |
| ✅ | Certified Tester Foundation Level (ISTQB) | ISTQB | 2022 |
