At InverseAI, I own the architecture of the core video-editing platform and re-architected its native decode and encode pipeline to offload processing to the GPU, increasing throughput by approximately 50%.
I expanded the editing pipeline from single-video to multi-video composition and engineered Filter and Adjust features with OpenGL and FFmpeg. GPU pipeline optimizations enabled faster rendering for 2K resolution and 60 fps videos.
On VidArt, I architected a native rendering engine in C++ and OpenGL ES / MediaCodec for real-time GPU playback and hardware-accelerated export. I also reworked the editor into a reusable module powering five InverseAI apps.
Across InverseAI’s Android apps, I’ve built video-processing systems and editing features, including an FFmpeg video-merging engine and a synchronized collage engine supporting up to nine concurrent video streams. I’ve also led a team of up to six engineers and mentored and onboarded engineers through code reviews and engineering standards.
