Bima S.

I'm an AI Engineer specialized in advanced computer vision and scalable AI deployments. My work spans real-time surveillance systems, traffic violation detection, and complex visual intelligence pipelines. I architect end-to-end AI systems from model development to backend deployment, integrating cutting-edge technologies like FaceNet, MediaPipe, dlib for face detection, recognition, and matching. My model optimizations include SPDConv, TorchScript, TensorRT, and CUDA acceleration on NVIDIA GPUs for ultra-fast inference. I'm experienced with YOLOv8, CSRNet for object counting (e.g., palm trees, vehicles), and integrate AI models into production-grade environments using FastAPI and Docker. I work with vector databases like Pinecone and implement Retrieval-Augmented Generation (RAG) systems using Langchain, LlamaIndex, and LLaMA vision models for advanced multimodal AI solutions. Strong focus on architecture design, GPU optimization, model quantization, and scalable vector search for high-performance AI applications. Passionate about pushing AI from research to real-world deployment, fully capable of remote collaboration and delivering production-ready solutions.