Machine & Robot Learning Engineer · Bengaluru, India
I build data and model flywheels for robot learning at Clutterbot. My work also includes training and deploying policies on the edge.
Both the models ran once on this photo, offline. The page does not run a model, and the errors stay visible.
Newest first.
May 2025 to present
Data pipelines and policy training for Robot Learning.
Jul 2024 to May 2025
As a part of my Bachelor’s Thesis, I built and benchmarked a Video Vision Transformer (ViViT) for classroom activity recognition. Used saliency maps to inspect the regions associated with model predictions. With Dr. Yuvaraj and Dr. Amalin, also worked on gaze estimation for student-engagement measures.
Jun 2024 to Aug 2024
Built a PyQt5 based annotation tool around Segment Anything (SAM) for AI-assisted image annotation. Automated the stitching of segmented regions into larger sub-scenes for scene understanding.
May 2023 to Aug 2023
Segmented and restored deteriorated Rajasthani murals under Dr. Dhiraj Sangwan. Compared U-Net++, DeepLabV3+, PSPNet and FPN for damage segmentation, and generated synthetic damage with StyleGAN2-ADA.
Tools and research projects, newest first.
2026
Built a tool that checks ONNX models for TensorRT compatibility before engine compilation, with static checks that need no GPU.
Feb 2024 to May 2024
Extended CloSe-Net for 3D clothing segmentation from coloured point clouds, with sharper edges between garments and automatic clothing-type detection.
Oct 2022 to Aug 2023
Smart glasses for visually impaired people, with TinyML object detection on ESP microcontrollers and an Android companion app.
Sep 2022 to Jan 2023
Worked on autonomous navigation for the BITS Goa student Mars rover: ROS path planning with A*, Dijkstra and SLAM, run in Gazebo on an NVIDIA Jetson Xavier.
Research work.
HS Baath, S. Shinde, J. Keniya, PR. Mishra, A. Saini, D. Sangwan
CVIP 2023 · Springer · Published 2024
Segments the damaged regions of a mural, then inpaints them. On its test images, the paper reports 0.892 mean intersection over union (mIoU) for segmentation. It reports 0.9812 structural similarity (SSIM) for inpainting.
Soham Shinde, Vikram Jamwal
ECCV AI4VA Workshop
Groups the segmented regions of a painting into larger sub-scenes for scene understanding.
From the CVIP 2023 paper
Move the slider to compare the damaged image with the model output. Select “Damage mask” to show the predicted damaged regions.
The model fills damaged areas with generated content. It does not recover the exact original painting. CSIR-CEERI, 2023.
B.E. Electronics & Communication Engineering
Minor in Data Science · Graduated July 2025
Email me about robotics, computer vision, or ML engineering queries or roles.
GitHub (opens in a new tab)LinkedIn (opens in a new tab)View CV (opens in a new tab)