AI vision platform for real-time retail and out-of-home advertising analytics
SpaceVision AI builds computer vision and behavior analytics infrastructure for digital out-of-home (DOOH) and retail environments. The stack—Python, TypeScript, Kafka, Kinesis, BigQuery, dbt, plus edge devices (ESP32, Raspberry Pi)—reveals a company processing real-time vision data at scale, with heavy investment in data pipeline orchestration (Airflow, dbt, Athena) and visualization layers (Tableau, Grafana, Streamlit). Pain points around scaling data pipelines for millions of records and real-time data flow, combined with active work on computer vision model training and an in-house annotation tool, suggest SpaceVision is building both the hardware-software bridge and the analytics layer in parallel.
SpaceVision AI, founded in 2023 and based in Seoul, operates a platform that applies machine learning and computer vision to capture and analyze audience behavior in retail and DOOH settings, converting attention and engagement into actionable advertising insights. The company is early-stage (2–10 employees) with engineering-led structure (4 engineers, 1 data, 1 product) and senior-skewed seniority distribution. Current roadmap spans intelligent ad CMS and player systems, real-time analytics dashboards, a global partner API, and an ad performance measurement tool. Active expansion challenges include scaling the data pipeline globally and entering EMEA and APAC markets.
Python, TypeScript, PostgreSQL, Next.js, NestJS, Kafka, Kinesis, BigQuery, Apache Airflow, dbt, Tableau, Grafana, and edge hardware (ESP32, Raspberry Pi). Cloud infrastructure runs on AWS and GCP.
AI analytics dashboards, intelligent ad CMS and player systems, real-time interactive analytics, a global platform API, ad performance measurement tools, and computer vision models trained on retail and DOOH data. Major focus on scaling data pipelines and expanding into EMEA and APAC markets.
Other companies in the same industry, closest in size