GStreamer Engineer (Remote / Offshore)
Agot is a mid-sized artificial intelligence (AI) startup based in Pittsburgh. We're building a camera AI product that will change the future of work. Agot Retina provides labor support and training, reduces waste, and "spellchecks" errors in real time.
As a GStreamer Engineer, you will be designing, building, and maintaining efficient video streaming pipelines using GStreamer for NVIDIA Jetson devices. You'll integrate computer vision models, optimize pipelines for hardware acceleration, and troubleshoot streaming issues. Strong coding skills (Python, C/C++), familiarity with media codecs and protocols, and experience with edge device deployment are essential. Collaboration, problem-solving, and communication skills are key in this role.
- Design, build, and maintain efficient, reusable, and reliable GStreamer pipelines for streaming video data.
- Develop video streaming applications using GStreamer, specifically targeting Nvidia Jetson devices.
- Integrate computer vision models into a GStreamer pipeline to enable real-time processing on edge devices.
- Optimize GStreamer pipelines to make full use of Jetson's hardware acceleration capabilities (CUDA, cuDNN, TensorRT).
- Debug and solve video streaming issues, like latency, quality, bandwidth usage, etc.
- Write well-structured, efficient code, ensuring high performance and responsiveness on edge devices.
- Document technical aspects of the developed applications and pipelines.
- Collaborate with cross-functional teams to define, design, and ship new features.
- Proven experience with the GStreamer framework and its integration with computer vision technologies. Experience with NVIDIA's deepstreamer SDK is a plus.
- Deep understanding of media codecs, formats, transports, and container protocols, such as H.264, H.265, VP8, VP9, AAC, RTMP, HLS, MPEG-DASH, and MP4.
- You will writing a library codebase. As such, familiarity with software design principles such as SOLID, DRY (Don't Repeat Yourself), and KISS (Keep It Simple, Stupid) is essential, as well as knowledge of design patterns and their implementation in real-world situations, which are key to creating maintainable and reusable code.
- Experience with video processing: encoding, decoding, transcoding, scaling, and familiarity with color spaces.
- Familiarity with NVIDIA Jetson's hardware acceleration (CUDA, cuDNN, TensorRT) in the context of GStreamer pipelines.
- Strong proficiency in Python, competency in C/C++
- Experience with real-time systems, multi-threading, and inter-process communication.
- Familiarity with Git, unit testing, debugging, profiling, system design, and issue tracking.
- Knowledge of network protocols like TCP/IP, UDP, and HTTP.
- Experience with deploying AI/ML models on edge devices, particularly NVIDIA Jetson series.
- Experience with library or API design for usability, performance, and maintainability.
- Experience in agile methodologies and software development lifecycles.
- Good problem-solving skills, excellent communication skills, and a team player attitude.
- A minimum of 3 years of professional experience in a related field.
- Previous experience working with NVIDIA Jetson devices and deploying models on the edge is strongly preferred.