Sakethram Madhuvarasu
Perception Software Engineer at Seneca
Hello all! I am Saketh, a Perception Software Engineer at Seneca, where I build the vision stack for autonomous drones — automated annotation and training pipelines, custom depth models, and LiDAR-based detection running on the edge.
Previously, I was a Machine Learning Engineer in Computer Vision at Meta Reality Labs, where my main focus was 3D reconstruction — generating synthetic training data from multi-view AR/VR captures using Gaussian Splatting — and a language-grounded 3D-VLM for 3D scene understanding and geometry-aware panel placement. I also worked on stereo depth estimation and hand–object segmentation. Before that, I worked on geometric computer vision at Vimaan Robotics — transformer-based detection and camera calibration for warehouse automation.
I earned my MS in Electrical and Computer Engineering at UC San Diego, specializing in Robotics and AI, and my B.Tech in Electrical Engineering at IIT Tirupati. At UCSD I was a graduate researcher on 3D Gaussian Splatting with Prof. Tara Javidi, and on visual-inertial SLAM and semantic odometry with Prof. Nikolay Atanasov.
I’m interested in 3D vision for robotics: 3D-scene reconstruction, multi-view geometry, 3D-perception, and Embodied AI
News
| May 2026 | Started as a Perception Software Engineer at Seneca |
|---|---|
| Mar 2026 | Our paper “Ego-1K: A Large-Scale Multiview Video Dataset for Egocentric Vision” was accepted at CVPR 2026 |
| Feb 2026 | Our paper “3D Scene Rendering with Multimodal Gaussian Splatting” was accepted at ACSSC 2026 |
| Aug 2025 | Started as a Machine Learning Engineer, Computer Vision at Meta Reality Labs |
| Jun 2025 | Graduated with an MS in ECE from UC San Diego |