Alok Kumar

I’m a Computer Vision and AI Engineer with hands-on experience building and deploying real-world deep learning and backend systems. I specialize in developing YOLO-based object detection models, LLM-powered applications (RAG, LangChain), and scalable microservices using Python, FastAPI, Docker, and Azure.

At DeepNet Analytics, I engineered real-time surveillance systems with optimized YOLOv8 models, boosting inference speed by 3× and deploying them across Azure ML Studio and edge devices (Jetson, VPS). I also designed event filtering algorithms to reduce false alerts by 80%, demonstrating my strength in both algorithmic optimization and system design.

Previously, at Lawgicx, I built a Retrieval-Augmented Generation (RAG) backend that enabled LLMs to answer legal queries from custom datasets — including features like persistent chat sessions, PDF parsing, and usage tracking dashboards.

My personal project, Anuwad, transforms PDFs and articles into AI-powered interactive learning experiences, integrating TTS, translation, summarization, and conversational AI — all within a Flutter-based app.

I’m passionate about AI system design, Generative AI, and real-time computer vision. I love solving technical challenges end-to-end — from model training and optimization to deployment and full-stack integration.