I'm an AI Engineer passionate about Deep Learning & Computer Vision. Building intelligent, agentic systems that solve real-world problems.

• Deep Learning: CNNs, RNNs, Transformers
• Computer Vision: Object detection, image classification, visual recognition
• Agentic AI: LLM tool-calling, policy-driven decision systems, audited autonomous agents
• Full-Stack AI: End-to-end AI solutions with modern web tech
• Machine Learning: Training models for real-world applications

Real-time PPE-compliance monitoring with YOLOv8 that not only detects violations but acts on them through an autonomous safety-supervisor agent.

• Tech: Python, YOLOv8, FastAPI, WebSockets, React, SQLAlchemy/SQLite, LLM (Groq / Gemini)
• Perception: Dual-engine detection (89.1% mAP) with custom spatial post-processing — anatomical PPE-to-person validation + nested-box IoSA NMS
• Agency: Policy-as-code escalation state machine, fully-audited incident trail, an LLM analytics assistant (function-calling + SQL tools), supervisor webhooks, and one-click PDF incident reports
• Impact: Turns passive detection into autonomous, accountable safety supervision

Deep Learning pipeline identifying 38 plant disease classes from leaf images.

• Tech: Python, TensorFlow, Keras, CNN
• Models: Custom CNN (96% accuracy) vs MobileNetV2 (90% accuracy)
• Dataset: 87K+ images with rigorous data isolation
• Impact: Early disease detection for farmers

Machine learning pipeline to predict property values using big data frameworks.

• Tech: Python, PySpark, Gradient-Boosted Trees
• Features: Custom feature engineering, scalable pipeline, rigorous RMSE and R² analysis
• Impact: Accurate real estate forecasting utilizing large-scale dataset processing

• Agentic AI & LLM orchestration
• MLOps & Model Deployment

• Full-time / entry-level opportunities in AI/ML & Computer Vision Engineering
• Internship opportunities in AI/ML Engineering
• Collaborating on AI/ML and Computer Vision projects
• Research opportunities in Deep Learning and CV

Made with ❤️ by @mamgad235