Mehak M

AI ENGINEER

Visibility Specialists • Fulltime

May 2025 - May 2026 • 1 yr

Worked on end-to-end AI systems including Computer Vision, NLP, Deep Learning, Agentic AI, and RAG-based architectures, covering full lifecycle from data collection to deployment and backend integration. Key Work Areas: 1. Data & ML Pipeline Collected and curated datasets from existing public sources Structured datasets into CSV format for training purposes Performed data cleaning, preprocessing, and feature engineering Built and managed end-to-end data pipelines for ML and computer vision models Prepared and optimized datasets for model training and evaluation 2. Computer Vision Built real-time YOLO-based detection systems (YOLOv8/YOLOv11) Developed solutions for intrusion detection, people counting/tracking, and fire/smoke detection Optimized models for edge devices like Raspberry Pi 3. Deep Learning Trained and fine-tuned models on GPU environments (Kaggle/P100) Improved performance using augmentation and hyperparameter tuning Evaluated models using precision, recall, and inference speed 4. RAG Systems Built multimodal RAG chatbot (text + image inputs) Implemented vector database search and embedding-based retrieval Integrated LLMs for context-aware and grounded responses 5.Agentic AI Designed multi-agent systems with a super-agent controller Implemented intent classification for task routing Used LangChain/LangGraph for workflow orchestration Deployed systems on Hugging Face for real-time inference 6.Backend & Deployment Built FastAPIs for AI integration Deployed ML models on Hugging Face Integrated AI systems into full-stack applications