name: Aryan Dev Tyagi
based_in: Ghaziabad, India
education: B.Tech CS (AI & ML) @ KIET Deemed University Β· CGPA 8.0 Β· 2023β2027
focus: Applied AI/ML Systems Β· GenAI Applications Β· AWS Serverless & Cloud Security
certified: AWS Solutions Architect (Associate) Β· AWS Cloud Practitioner Β· AWS AI Practitioner
building: Telecom NOC AI β multi-model fault prediction + GenAI NOC copilot
HireMind AI β an AI-powered mock interview platform
CloudGuard β a serverless AWS security scanner
leads: President, KIET Quizzinga Club
fun_fact: IIT Delhi Rendezvous Quiz Finalist Β· UGMC Delhi Runner-up- π€ I build end-to-end AI/ML systems β ensemble classifiers, deep learning, anomaly detection, and GenAI copilots wired into real full-stack apps, not just notebooks
- βοΈ I design and ship serverless, least-privilege AWS architectures β Lambda, API Gateway, S3, IAM, DynamoDB
- π‘οΈ I care about cloud security β CloudGuard automatically catches IAM misconfigurations & S3 exposure
- π§ I'm the President of KIET Quizzinga Club, running quiz events for 80+ participants with a 35+ member team
- β‘ Off-screen: cricket & football (Chelsea π, Mumbai Indians π), history, and retro cinema
| May 2026 β Jun 2026 |
Web Development Intern Β· Yinolite Solution Built a responsive JavaScript Quiz Application β quiz management, timer, scoring, result generation, and REST API integration. Collaborated via Git/GitHub and improved frontend responsiveness across devices. |
| Dec 2025 β Feb 2026 |
AWS Cloud & Generative AI Virtual Intern Β· SmartInternz (SmartBridge) Built a serverless app on AWS Lambda, API Gateway & S3, and configured IAM least-privilege policies β resolving 10+ permission issues during a Generative AIβfocused capstone. |
Python Β· FastAPI Β· React Β· TypeScript Β· PyTorch Β· XGBoost Β· scikit-learn Β· Google Gemini Β· SQLite Β· JWT
An AI-powered Network Operations Center decision-support platform, built on the real Kaggle Telstra Network Disruptions dataset. Not just a classifier β a full detect β predict β explain β diagnose β prevent β remediate pipeline for NOC engineers.
- π§© 5-model ensemble (Random Forest, XGBoost, SVM, LSTM, GRU) for fault severity prediction, F1-weighted soft voting β best individual model (XGBoost) at 72.71% accuracy
- π PyTorch autoencoder trained on normal-only telemetry for independent anomaly detection, fused into one combined operational risk score
- π©» Explainable AI feature attribution + a Gemini-powered NOC Copilot (root-cause hypotheses, preventive recommendations, remediation plans) with a fully offline deterministic fallback
- π JWT-authenticated dashboard, deployed full-stack (Render + Vercel)
- ποΈ Includes a live What-If Network Simulator β test hypothetical scenarios through the same real ML engine
|
React Β· Node.js Β· Express Β· MongoDB Β· GenAI Β· JWT Β· Razorpay AI-powered mock interview platform generating role-specific questions with a secure, scalable backend.
|
AWS Lambda Β· API Gateway Β· S3 Β· IAM Β· DynamoDB Β· React Β· Recharts Serverless cloud security scanner that detects IAM misconfigurations and S3 vulnerabilities in near real time.
|
|
Flask Β· React Β· AWS EC2 Β· DynamoDB Β· JWT Real-time cryptocurrency analytics platform deployed on AWS.
|
Python A movie recommendation system. |
CollabB β AI Moderation & Coaching Platform Contributed through issues, feature branches, and PRs in a multi-contributor codebase. Authored a merged technical proposal covering Unicode Normalization (NFKC), text sanitization, prompt-injection prevention, and input validation for AI-safety-conscious moderation workflows.
| π IIT Delhi Rendezvous Quiz Finalist | π₯ UGMC Delhi Quiz Runner-up |
| π Chelsea π & Mumbai Indians π fan | ποΈ Retro music & classic film buff |
| π History lover & lifelong learner | π§ President, KIET Quizzinga Club |