I am an Assistant Professor in Computer Science at the University of Toronto. Prior to coming to Toronto, I was a visiting Research Scientist at TRI. I was also a Research Scientist at MIT as part of the Computer Science and Artificial Intelligence Lab working with Sertac Karaman and Daniela Rus.

I joined MIT from the Autonomous Systems Lab of ETH Zurich where I worked with Roland Siegwart on robotic perception and navigation. I obtained my PhD degree in computer science working on nonlinear estimation at the Intelligent Sensor-Actuator-Systems Lab of the Karlsruhe Institute of Technology. It was supervised by Uwe Hanebeck and Simon Julier (co-supervisor).

Early on, I was fascinated by the laws that govern uncertainty and their applications, which was my focus while studying Mathematics (Major) and Computer Science (Minor) at the University of Stuttgart. My work aims at enabling robust interactive autonomy for robotics by developing novel perception and decision making methods for challenging dynamic environments.

- Dissecting Deep RL with High Update Ratios: Combatting Value Divergence
- When does Self-Prediction help? Understanding Auxiliary Tasks in Reinforcement Learning
- LEOD: Label-Efficient Object Detection for Event Cameras
- Producing and Leveraging Online Map Uncertainty in Trajectory Prediction
- SPAD: Spatially Aware Multi-View Diffusers
- AvatarOne: Monocular 3D Human Animation
- iNVS: Repurposing Diffusion Inpainters for Novel View Synthesis
- SlotDiffusion: Object-Centric Generative Modeling with Diffusion Models
- trajdata: A Uniļ¬ed Interface to Multiple Human Trajectory Datasets
- Geometry Matching for Multi-Embodiment Grasping
- Dynamic Multi-Team Racing: Competitive Driving on 1/10-th Scale Vehicles via Learning in Simulation
- Reference-guided Controllable Inpainting of Neural Radiance Fields
- GROUNDED: A Localizing Ground Penetrating Radar Evaluation Dataset for Learning to Localize in Inclement Weather
- Multi-Abstractive Neural Controller: An Efficient Hierarchical Control Architecture for Interactive Driving
- Invertible Neural Skinning
- SparsePose: Sparse-View Camera Pose Regression and Refinement
- SPIn-NeRF: Multiview Segmentation and Perceptual Inpainting with Neural Radiance Fields
- Solving Continuous Control via Q-Learning
- Housekeep: Tidying Virtual Households using Commonsense Reasoning
- LaTeRF: Label and Text Driven Object Radiance Fields
- MapLite 2.0: Online HD Map Inference Using a Prior SD Map
- Neighborhood Mixup Experience Replay: Local Convex Interpolation for Improved Sample Efficiency in Continuous Control Tasks
- A Deep Concept Graph Network for Interaction-Aware Trajectory Prediction
- HYPER: Learned Hybrid Trajectory Prediction via Factored Inference and Adaptive Sampling
- Learning Interactive Driving Policies via Data-driven Simulation
- VISTA 2.0: An Open, Data-driven Simulator for Multimodal Sensing and Policy Learning for Autonomous Vehicles
- Learning An Explainable Trajectory Generator Using The Automaton Generative Network (AGN)
- Sensitivity-Informed Provable Pruning of Neural Networks
- Is Bang-Bang Control All You Need? Solving Continuous Control with Bernoulli Policies
- Strength Through Diversity: Robust Behavior Learning via Mixture Policies
- GROUNDED: The Localizing Ground Penetrating Radar Evaluation Dataset
- Vehicle Trajectory Prediction Using Generative Adversarial Network with Temporal Logic Syntax Tree Features
- Deep Latent Competition: Learning to Race Using Visual Control Policies in Latent Space
- Differentiable Logic Layer for Rule Guided Trajectory Prediction
- Exploiting Semantic and Public Prior Information in MonoSLAM
- Deep Context Maps: Agent Trajectory Prediction using Location-specific Latent Maps
- Deep Orientation Uncertainty Learning based on a Bingham Loss
- Autonomous Navigation in Inclement Weather based on a Localizing Ground Penetrating Radar
- Learning Robust Control Policies for End-to-End Autonomous Driving from Data-Driven Simulation
- MapLite: Autonomous Intersection Navigation without a Detailed Prior Map
- Infrastructure-free NLoS Obstacle Detection for Autonomous Cars
- Range-based Cooperative Localization with Nonlinear Observability Analysis
- Probabilistic Risk Metrics for Navigating Occluded Intersections
- Appearance-Based Landmark Selection for Visual Localization
- Data-Dependent Coresets for Compressing Neural Networks with Applications to Generalization Bounds
- Directional Statistics and Filtering Using libDirectional
- Inferring Pedestrian Motions at Urban Crosswalks
- The Voliro Omniorientational Hexacopter: An Agile and Maneuverable Tiltable-Rotor Aerial Vehicle
- ShadowCam: Real-Time Detection of Moving Obstacles behind a Corner for Autonomous Vehicles
- LandmarkBoost: Efficient Visual Context Classifiers for Robust Localization
- Why and How to Avoid the Flipped Quaternion Multiplication
- Free LSD: Prior-Free Visual Landing Site Detection for Autonomous Planes
- Incremental Segment-Based Localization in 3D Point Clouds
- maplab: An Open Framework for Research in Visual-inertial Mapping and Localization
- Map Management for Efficient Long-Term Visual Localization in Outdoor Environments
- Sampling-Based Approximation Algorithms for Reachability Analysis with Provable Guarantees
- A Low-Cost System for High-Rate, High-Accuracy Temporal Calibration for LIDARs and Cameras
- Onboard Real-time Dense Reconstruction of Large-scale Environments for UAV
- Evaluation of Combined Time-Offset Estimation and Hand-Eye Calibration on Robotic Datasets
- Efficient Descriptor Learning for Large Scale Localization
- Map Quality Evaluation for Visual Localization
- TSDF-Based Change Detection for Consistent Long-Term Dense Reconstruction and Dynamic Object Discovery
- Visual-Inertial Self-Calibration on Informative Motion Segments
- Gone with the Wind: Nonlinear Guidance for Small Fixed-Wing Aircrafts in Arbitrarily Strong Windfields
- Robust Estimation and Applications in Robotics
- Monocular Visual-Inertial SLAM for Fixed-Wing UAVs Using Sliding Window Based Nonlinear Optimization
- A Data-Driven Approach for Pedestrian Intention Estimation
- Will it last? Learning Stable Features for Long-Term Visual Localization
- Appearance-Based Landmark Selection for Efficient Long-Term Visual Localization
- Erasing Bad Memories: Agent-Side Summarization for Long-Term Mapping
- Generalized Information Filtering for MAV Parameter Estimation
- Robust Map Generation for Fixed-Wing UAVs with Low-Cost Highly-Oblique Monocular Cameras
- Collaborative 3D Reconstruction using Heterogeneous UAVs: System and Experiments
- Optimal Quantization of Circular Distributions
- Non-Parametric Extrinsic and Intrinsic Calibration of Visual-Inertial Sensor Systems
- Unscented von Mises-Fisher Filtering
- Recursive Bayesian Filtering in Circular State Spaces
- Unscented Orientation Estimation Based on the Bingham Distribution
- Exact and Approximate Hidden Markov Chain Filters Based on Discrete Observations
- A Stochastic Filter for Planar Rigid-Body Motions
- Non-Identity Measurement Models for Orientation Estimation Based on Directional Statistics
- Bivariate Angular Estimation Under Consideration of Dependencies Using Directional Statistics
- Recursive Bingham Filter for Directional Estimation Involving 180 Degree Symmetry
- A Direct Method for Checking Overlap of Two Hyperellipsoids
- Efficient Evaluation of the Probability Density Function of a Wrapped Normal Distribution
- Efficient Bingham Filtering based on Saddlepoint Approximations
- The Partially Wrapped Normal Distribution for SE(2) Estimation
- A New Probability Distribution for Simultaneous Representation of Uncertain Position and Orientation
- Deterministic Approximation of Circular Densities with Symmetric Dirac Mixtures Based on Two Circular Moments
- Deterministic Dirac Mixture Approximation of Gaussian Mixtures
- Nonlinear Measurement Update for Estimation of Angular Systems Based on Circular Distributions
- Bearings-Only Sensor Scheduling Using Circular Statistics
- Recursive Estimation of Orientation Based on the Bingham Distribution
- Efficient Deterministic Dirac Mixture Approximation
- Recursive Nonlinear Filtering for Angular Data Based on Circular Distributions
- A Robust Computational Test for Overlap of Two Arbitrary-dimensional Ellipsoids in Fault-Detection of Kalman Filters