Preventai is a fast-growing Swedish AI company transforming how surveillance infrastructure is used in retail. Our platform connects to existing camera systems and applies advanced Video Language Models (VLMs) to detect incidents and generate real-time alerts. We work with some of Sweden's largest grocery chains and are expanding our team to support increasingly complex networked environments.
About the Role
We are looking for an experienced (at least 2-4 years of experience) Machine Learning / Computer Vision Engineer to build and deploy real-time AI systems focused on video analytics. You will work on developing, optimizing, and deploying tracking, action recognition, and VLM-based models in production environments. Strong problem-solving skills and a deep understanding of scalable inference pipelines are essential.
Responsibilities
- Develop, optimize, and deploy Computer Vision, Tracking, Action Recognition, and VLM models to production.
- Build and manage real-time video pipelines using RTSP and FFmpeg for multi-stream handling.
- Deploy AI inference services on GCP / Cloud / local server infrastructure with auto-scaling and GPU acceleration.
- Build scalable inference services using Flask / FastAPI, Docker, and REST APIs.
- Work with CUDA, GPU optimization, TensorRT/ONNX for real-time model performance.
- Implement model optimization techniques (quantization, pruning, batching, async pipelines).
- Build robust multi-threaded / multi-process systems for high-throughput streaming.
- Debug and profile performance bottlenecks across model, networking, and system layers.
- Collaborate across research and engineering teams to deliver production-ready solutions.
Required Skills
- Strong programming skills in Python.
- Hands-on experience with model deployment in real-time applications.
- Expertise in RTSP streaming pipelines (decoding, multi-camera ingest, latency optimization).
- Strong experience with FFmpeg for real-time video manipulation.
- Cloud deployment experience with GCP (firestore / Cloud Run / GPU instances / Storage).
- Strong knowledge of CUDA, GPU acceleration, memory optimization.
- Experience building APIs using Flask / FastAPI, and creating robust REST API services.
- Proficiency with Docker and container-based deployments.
- Strong debugging, profiling, performance optimization experience.
- Experience with multithreading / multiprocessing.
- Experience working with tracking models, action recognition models, or VLMs.
Nice to Have
- Experience with TensorRT, ONNX Runtime, Triton Inference Server, TorchServe.
- Real-time video streaming technologies (RTSP,HLS, WebRTC).
- CI/CD automation and DevOps best practices.
What We Offer
- Work in a high-impact role within a Swedish funded AI company
- Direct influence on how cutting-edge AI systems are deployed across real-world infrastructures
- Access to a highly technical team with expertise in AI, computer vision, and systems engineering
- Competitive compensation and benefits
- Remote-first setup with occasional on-site fieldwork
- Opportunity to define and scale infrastructure practices from the ground up
How to Apply
Please apply through LinkedIn, and please contact us at Preventai through LinkedIn in case you have any questions.
