ArguSight AI

ArguSight AI helps product and engineering teams turn image, video, and document problems into evaluated, integrated software. Projects can span technical feasibility and proof of concept through deployment, validation, and handoff. ArguSight is a founder-led computer vision and applied AI company. Clients work directly with Theodoros Papaiakovou—the engineer responsible for scoping and delivery—so technical decisions stay close to implementation. Core capabilities: • Computer vision and video analytics: object detection, tracking, segmentation, event detection, and temporal analysis • 2D/3D vision and human movement: pose estimation, camera geometry, reconstruction, biomechanics, scoring, and feedback • OCR and document AI: structured extraction, annotation and review tools, and spreadsheet or database workflows • Generative imaging and production ML: model fine-tuning, diffusion and GAN workflows, GPU inference, and deployment • AI integration: APIs, backend systems, tool-using agents, human approvals, evaluation, logging, and guardrails Engagements can begin with technical discovery, a data review, a focused prototype, or improvements to an existing system. Delivery can include model development and evaluation, backend/API integration, ONNX/TensorRT optimization, Docker-based deployment, testing, documentation, and handoff. Selected work includes a 3D motion-analysis and coaching pipeline, customer-facing generative image and video visualization tools, a human-in-the-loop OCR workflow for industrial drawings, and an AI-assisted Django/PostgreSQL workflow with explicit human escalation. Typical stack: Python, C++, PyTorch, TensorFlow, OpenCV, CUDA, ONNX, TensorRT, Django, PostgreSQL, Docker, AWS, and GCP. ArguSight AI is led by Theodoros Papaiakovou, an applied AI and computer vision engineer with an M.S. in Electrical and Computer Engineering from UCLA. ArguSight AI LLC is a Florida-based company working with clients internationally. Have a computer vision or AI workflow problem? Send an Upwork message with the use case, available data, desired output, and intended deployment environment.