Ashish R.

I’m an AI & Computer Vision Engineer specializing in building, training, optimizing, and deploying real-world AI solutions for image and video applications. I work across the complete computer vision pipeline — from dataset preparation and model training to inference, object tracking, Re-ID, optimization, and production deployment. If you need someone to build a computer vision solution from scratch, improve an existing model, or integrate AI into your production system, I can help. 🚀 What I Can Build 🔹 Object Detection - YOLO-based object detection - Custom object detection models - Small-object detection - Custom dataset preparation and training - Detection in images, videos, CCTV and RTSP streams 🔹 Object Tracking - Multi-object tracking - Object detection + tracking pipelines - Entry/exit tracking and counting - Appearance-based tracking - Re-ID based tracking - Tracking objects across occlusions and difficult scenes 🔹 Object Re-Identification - Feature/embedding extraction - Similarity matching - Cosine similarity based matching - Appearance-based identity association - Re-ID integrated with detection and tracking 🔹 Image Segmentation - Instance segmentation - Semantic segmentation - Object-level segmentation - Detection + segmentation + tracking pipelines 🔹 Image Classification - Custom classification models - Binary and multi-class classification - Model evaluation and threshold optimization - Production-ready classification pipelines 🔹 Video Analytics - Real-time video processing - RTSP/CCTV analytics - Multi-camera processing - FPS and frame-drop monitoring - Object counting and zone-based analytics - Entry/exit detection - Real-time event detection ⚙️ Model Optimization & Deployment I can take a trained deep learning model and optimize it for real-world inference and deployment. Experience includes: PyTorch → ONNX → OpenVINO / TensorRT / TFLite I work on: - GPU inference optimization - CPU inference optimization - ONNX conversion - OpenVINO deployment - TensorRT optimization - TFLite conversion - Edge AI deployment - Model size and latency optimization - Production inference pipelines 🧠 Deep Learning & AI My experience includes: - PyTorch - TensorFlow - YOLO / Ultralytics - OpenCV - ONNX - OpenVINO - TensorRT - TFLite - CNNs - Transformers - Embeddings - Metric learning - Re-identification - Image classification - Object detection - Image segmentation - Pose estimation 🎥 Real-Time Computer Vision I have experience working with real-time video pipelines where performance and reliability matter. I can work with: - RTSP streams - IP cameras - CCTV systems - MP4/video files - Multiple simultaneous streams - CPU/GPU inference - Real-time detection and tracking I can also help identify production issues such as FPS drops, frame loss, inference bottlenecks, synchronization problems, memory usage and latency. 🤖 NLP & AI Assistants Along with computer vision, I have experience developing NLP-based chatbots and AI assistants that can understand user queries and provide relevant responses. 🛠️ My Development Approach I don't focus only on getting a model to work in a notebook. My goal is to build a solution that works reliably in the real-world environment. Typical workflow: Requirement → Dataset → Training → Evaluation → Optimization → Deployment → Production Testing I can work with an existing codebase/model or develop the complete solution from scratch. 💼 Projects I Can Help With Whether you need: - A custom YOLO model - Object detection for a specific product or machine - Real-time CCTV analytics - Object tracking and counting - Re-ID based tracking - Image/video segmentation - Model conversion to ONNX/OpenVINO/TensorRT - Edge AI deployment - AI-powered inspection systems - Computer vision automation - Video stream monitoring - Existing model optimization - Debugging an AI production pipeline I can help design and implement the solution. ⭐ Why Work With Me? I combine computer vision, deep learning, model optimization, and deployment rather than focusing only on model training. I’m comfortable working with both research/prototyping and production-oriented computer vision systems, and I enjoy solving challenging problems involving images, videos, cameras, tracking, identity recognition and real-time AI. If you have a Computer Vision, AI, Machine Learning, Video Analytics, Object Detection, Tracking, Re-ID or Model Deployment project, feel free to message me with the requirements. I’d be happy to discuss the approach and help turn your idea into a working solution.