PhD Candidate · WPI Robotics Engineering · Worcester, MA

Venkatesh
Mullur

Seeking research and engineering roles in robotic manipulation, robot perception, and physical AI — where the goal is systems that deploy, not just papers that publish.

View Research Work Publications GitHub ↗

Venkatesh Mullur

Years in CV &
Robotics Research

Publications &
Conferences

Robot Platforms
Deployed On

🏊

National-level
competitive swimmer

Role Robotics Researcher

Lab MER Lab, WPI ↗ Advisor Prof. Berk Calli ↗

Status Open to opportunities

01 / About

Perception, grasping, and manipulation for robots that handle uncertainty

I'm a PhD candidate in Robotics Engineering at Worcester Polytechnic Institute, advised by Prof. Berk Calli. I work on robot perception, grasp detection, and vision-language systems for manipulation in cluttered and uncertain environments.

My background is in Electronics Engineering from Savitribai Phule Pune University, which gave me a strong foundation in embedded systems and low-level signal processing before I moved into robotics and deep learning at WPI. That path from hardware to software shapes how I think — I care about the full stack, not just the model.

I'm actively looking for roles in robotic manipulation, robot perception, and physical AI at companies building systems that need to work outside a controlled lab setting. Also affiliated with CogniAux, a faculty-founded computer vision startup.

What I can do from day one

⚙️

Deploy perception pipelines on real robots Franka Panda, Yale OpenHand, ROS2 — I've taken systems from training to hardware execution, not just simulation

🧠

Train and evaluate deep learning models end-to-end From dataset curation and architecture design to benchmark evaluation and ablations — PyTorch, CUDA, HPC clusters

📐

Design rigorous evaluations and benchmarks Not just best-case numbers — I quantify failure modes, diversity, and complementarity across datasets and conditions

📄

Contribute to both research and engineering Comfortable writing papers and shipping code — I've done both under deadline across three publications

02 / Research Projects

Research
Work

Flagship · MER Lab 2023 – 2024 IJRR Submission

Occlusion-Robust Robot Perception Pipeline

Real-time perception stack for encoderless Franka Panda. WGAN-GP inpainting reconstructs occluded joints from markerless images; keypoint detection with UKF temporal smoothing feeds Image-Based Visual Servoing, enabling stable control under severe occlusion. Outperforms LaMa with 11.7 FID, 0.91 precision.

WGAN-GP Keypoint RCNN GNN Tracking UKF IBVS Franka Panda PyTorch

91.6%

Occlusion accuracy

<2%

Pixel error

26%

Settling time ↓

Grasping Research 2024 – 2026 IROS 2026 Submission

Feature-Level Mixture-of-Experts for Grasping

Feature-level fusion across multiple grasp expert networks. Complementarity quantified via Q-statistics and error correlation — proving mid-accuracy moderately-correlated ensembles outperform SOTA individual models. Evaluated on Cornell, Jacquard, and GraspNet-1B; validated on Franka Panda.

Mixture-of-Experts Feature Fusion GR-ConvNet GG-CNN Q-Statistics GraspNet-1B PyTorch

13%

Fewer grasp failures

+7%

Over individual models

Physical AI 2026 – Present

Vision-Language-Action Robotic Grasping

Proposal-conditioned VLA framework for cluttered tabletop grasping with Yale OpenHand gripper. RGB-D Mask R-CNN generates object proposals; confidence-based skill selection chooses the best planner across known, unknown, and occluded multi-object scenarios. Deployed end-to-end on real hardware.

VLA Mask R-CNN RGB-D Skill Selection Yale OpenHand ROS2

92%

Grasp success in clutter

Industry · CogniAux 2024

Multimodal Human Activity Tracking

Real-time fusion of RGB-D, EEG, and audio for human activity tracking and attention estimation. Led a three-person team to full deployment: OpenPose, temporal attention, GMM, and optical flow as a full-stack GUI — not just a research prototype.

Sensor Fusion OpenPose Vision-Language GMM Optical Flow RGB-D + EEG

+17%

mAP improvement

43%

Error rate ↓

03 / Coursework & Side Projects

More
Projects

Coursework and independent projects spanning 3D vision, robotics systems, and deep learning.

3D Vision

Neural Radiance Fields

Enhanced NeRF with improved rendering and sampling strategies for photorealistic novel view synthesis.

PyTorchNeRF3D Rendering

View ↗

3D Reconstruction

Structure from Motion

Full SfM pipeline with epipolar geometry and bundle adjustment for monocular depth estimation.

OpenCVEpipolar GeometryBundle Adjustment

View ↗

3D Perception

Point Cloud Semantic Segmentation

3.2% IoU improvement via voxel grid filtering and bird's-eye-view representation with PointNet++.

PointNet++PCLOpen3D

View ↗

SLAM & Odometry

Visual-Inertial Odometry (MSCKF)

VIO at 29 FPS via stereo camera and IMU fusion for real-time 6-DOF pose estimation.

MSCKFIMU FusionROS

View ↗

Robotics

Motion Planning in Adversarial Environments

RRT-APF hybrid planner for robot navigation in dynamic environments with moving obstacles.

RRTAPFMotion Planning

View ↗

Computer Vision

Automatic Camera Calibration

Automated multi-view calibration pipeline with robust homography estimation and distortion correction.

OpenCVHomographyCamera Models

View ↗

04 / Skills

Technical
Depth

Robotics & Systems

ROS1 / ROS2
MoveIt! · RViz
Gazebo · Isaac Sim
Franka Panda
Yale OpenHand
Visual Servoing (IBVS)
Visual SLAM

Perception & Vision

RGB-D · Point Clouds
Keypoint Detection
Pose Estimation
Occlusion Handling
OpenCV · Open3D · PCL
Epipolar Geometry
Sensor Fusion

Deep Learning

PyTorch · TensorFlow
GANs · GNNs
Vision-Language Models
Mixture-of-Experts
Mask R-CNN
CUDA · HuggingFace
Few-shot Learning

Engineering

Python · C/C++
Docker · AWS
Git · CI/CD
Bash · SQL · MATLAB
HPC Clusters
Lidar / Radar
Kubernetes

05 / Research

Publications
& Research

Under Review

Feature-Level Mixture-of-Experts for Robust Robotic Grasp Detection

Venkatesh Mullur, Vinayak Kapoor, Prof. Berk Calli

IEEE IROS 2026 · Under review

Under Review

Utilizing Inpainting for Keypoint Detection for Vision-Based Control of Robotic Manipulators

Sreejani Chatterjee, Venkatesh Mullur, Abhinav Gandhi, Prof. Berk Calli

IJRR 2025 · Under review

Presented

Novel Sweeping Methods for Robotic Rearrangement of Object Piles

Abhijeet Sanjay Rathi, Filip Radil, Hrishikesh Dhairyasheel Pawar, Prof. Berk Calli

IEEE CASE 2025 · Los Angeles · August 2025 · Presenter

Let's build robots that survive the real world.

Actively seeking roles in robotic manipulation, robot perception, and deep learning for physical AI. If your system needs to work in clutter, occlusion, and real deployment conditions — let's talk.

vmullur@wpi.edu ↗ LinkedIn — @vmullur ↗ GitHub — venk221 ↗ +1 (508) 304-2824 ↗

Worcester, MA · Open to relocation · PhD exp. 2028

VenkateshMullur

Perception, grasping, and manipulation for robots that handle uncertainty

ResearchWork

Occlusion-Robust Robot Perception Pipeline

Feature-Level Mixture-of-Experts for Grasping

Vision-Language-Action Robotic Grasping

Multimodal Human Activity Tracking

MoreProjects

Neural Radiance Fields

Structure from Motion

Point Cloud Semantic Segmentation

Visual-Inertial Odometry (MSCKF)

Motion Planning in Adversarial Environments

Automatic Camera Calibration

TechnicalDepth

Publications& Research

Feature-Level Mixture-of-Experts for Robust Robotic Grasp Detection

Utilizing Inpainting for Keypoint Detection for Vision-Based Control of Robotic Manipulators

Novel Sweeping Methods for Robotic Rearrangement of Object Piles

Let's build robots that survive the real world.

Venkatesh
Mullur

Research
Work

More
Projects

Technical
Depth

Publications
& Research