AI-powered ground operations optimization for airports and airlines
Assaia deploys computer-vision and ML systems into airport operations, running inference on live video feeds (FFmpeg, GStreamer, PyTorch, Triton) to optimize aircraft turnarounds and gate utilization. The stack reveals a production-heavy focus: Kubernetes orchestration, PostgreSQL + MongoDB for state, Celery for task queues, and FastAPI for real-time APIs — all necessary to handle continuous video ingest and low-latency ML predictions at airport scale. Active hiring is engineering-concentrated (5 of 9 roles), with seniority skewed senior, suggesting they're building out inference infrastructure and MLOps maturity rather than expanding headcount broadly.
Assaia provides AI software for optimizing turnaround operations at airports and airlines. The platform ingests live CCTV feeds, applies neural-network inference to predict and automate ground tasks, and surfaces insights to airport operations and airline crews via web and mobile interfaces. Deployments span Europe and North America, serving major airports (London-Gatwick, Seattle-Tacoma, Rome, Toronto Pearson, JFK) and carriers including British Airways, United Airlines, and Alaska Airlines. Core pain points in the platform roadmap center on ML pipeline scalability, MLOps practices, and reducing technical debt—typical signals of a company moving from prototype toward repeatable, multi-tenant production deployments.
Python, PyTorch, Triton Inference Server, Kubernetes, Docker, PostgreSQL, MongoDB, Redis, FastAPI, Django, TypeScript/React. Video handling via FFmpeg and GStreamer; CI/CD through GitLab.
ML pipelines, real-time inference infrastructure, cloud-based CCTV streaming, MLOps improvements, AI-powered turnaround management, and internal design systems. Active focus on reducing technical debt and building scalable systems.
Other companies in the same industry, closest in size