Robotics · Embodied Intelligence · Machine Learning

Research in robot perception, navigation, and spatial intelligence.

I’m a senior Computer Engineering student at NYU Abu Dhabi, minoring in Applied Mathematics. My research focuses on multimodal sensing, visual localization, monocular geometry, semantic 3D scene understanding, and robust embodied navigation. I also work on efficient machine-learning systems and probabilistic modeling.

I plan to pursue PhD research in robotics and embodied intelligence after graduating in 2027.

3.988GPA / 4.000
2First-author papers
1stResearch Poster, NYUAD ’25
GoldiGEM 2025
Portrait of Aman Sunesh

Current work

Research

My current work centers on robust perception and navigation for embodied systems, with complementary work in trajectory forecasting and machine-learning systems.

Building Informatics and Visualization Lab (BiLAB) — Undergraduate Researcher

NYU Tandon School of Engineering

Jun 2026 – Present
  • Built a multimodal work-zone trajectory forecasting dataset with 26,892 driving sequences from 33 participants, integrating trajectories, surrounding agents, driver gaze, controls, work-zone geometry, and map context.
  • Developed a participant-disjoint K=6 forecasting benchmark spanning physics-based and state-of-the-art methods including LAFormer, MTR, Forecast-MAE, and QCNet.
  • Designed WZ-TARF, a work-zone-conditioned forecasting architecture achieving 1.038/2.010 m minADE/minFDE and improving over the strongest benchmarked baseline by 12.0%/21.2%.
Trajectory ForecastingPyTorchMultimodal LearningAutonomous Driving

Senior Design Capstone — Semantic 3D Mapping for Object-Goal Navigation

NYU Abu Dhabi

2026 – 2027
  • Developing an end-to-end object-navigation system combining 3D Gaussian Splatting SLAM, open-vocabulary semantic mapping, object-level spatial reasoning, and ROS 2 navigation.
  • Benchmarking mapping robustness under dynamic objects, occlusion, illumination changes, and sensor noise, with emphasis on downstream navigation reliability.
3D Gaussian SplattingSLAMSemantic MappingROS 2

Earlier research

Dynamical Systems Lab, NYU Tandon

Research Assistant · Feb – May 2026

Built a U.S.-wide tract-level migration–hazard pipeline over 157M directed flows and estimated Poisson gravity models across four hazard families with extensive spatial and robustness checks.

Center for Quantum & Topological Systems, NYU Abu Dhabi

Undergraduate Researcher · Sep 2024 – Sep 2025

Engineered an NV-center quantum diamond spectrometer and developed a PyQt6/QuPyT experiment-control and analysis platform for reproducible quantum-sensing workflows.

Peer-reviewed work

Publications

LOCO 2026 2026

RequestRouter: Request-Boundary Routing for Efficient Single-GPU LLM Inference

Aman Sunesh, Ali Alshehhi, Hivansh Dhakne

Request-boundary routing across fixed inference modes on a single NVIDIA A100, improving latency and energy efficiency while maintaining benchmark quality.

AALTD 2026 · ECML-PKDD 2026

Modeling Information Blackouts in Missing Not-At-Random Time Series Data

Aman Sunesh, Allan Ma, Siddarth Nilol

State-space modeling of traffic-sensor blackouts as informative missingness, separating blackout reconstruction from post-blackout forecasting and comparing against MAR and classical baselines.

Selected systems

Projects

A small selection of projects that best represent my work in robotics, perception, and hardware–software systems.

Current · Senior Capstone

Robust Semantic 3D Gaussian Splatting for Object-Goal Navigation

End-to-end robotic object navigation using 3DGS-based SLAM, semantic mapping, object-level spatial reasoning, and navigation evaluation under realistic scene changes.

3DGSSLAMEmbodied Navigation

Robot Vision

Autonomous Maze Navigation Using Visual-Geometric SLAM

Built a visual-geometric navigation system using RootSIFT-VLAD localization, SLAM-estimated poses, epipolar refinement, and direction-aware A* planning; completed the maze with 0.051 m final translation error.

Robotics

Autonomous Indoor Navigation with LiDAR and AprilTags

Hardware navigation stack combining 2D LiDAR occupancy mapping, wheel odometry, AprilTag localization, A* planning, wall centering, and emergency obstacle stopping.

Experimental sensing

NV-Center Quantum Sensing Platform

Integrated optics, RF/microwave control, DAQ, photon detection, and PyQt6/QuPyT software into a reproducible quantum diamond spectroscopy workflow.

Academic background

Education

New York University Abu Dhabi

B.S. Computer Engineering · Minor in Applied Mathematics

Aug 2023 – May 2027 GPA 3.988 / 4.000

Selected coursework

Space Robotics Graduate Control Systems Engineering Embedded Systems Applied Machine Learning Data Structures & Algorithms Mathematical Modeling

NYU Tandon School of Engineering

Study Away · Computer Engineering

Aug 2025 – May 2026 GPA 4.000 / 4.000

Selected coursework

Probabilistic Time Series Analysis Graduate Efficient AI & Hardware Accelerator Design Graduate Robotic Manipulation & Locomotion Robot Vision Artificial Intelligence

Selected recognition

Awards

1st

Best Research Poster

NYUAD Summer Research Symposium · 2025

Recognized for the NV-center quantum sensing and quantum diamond spectrometer project.

Gold

iGEM 2025

Gold Medal · Best Integrated Human Practices

Software and modeling lead for a CDG-focused synthetic biology project and PK/PD simulation platform.

IBM

Quantum Excellence Badge

IBM Quantum

Completed advanced Qiskit work in quantum programming, error suppression and mitigation, and utility-scale simulation.

Get in touch

Interested in robotics and embodied intelligence research.

I’m always happy to talk about research, collaborations, or PhD opportunities for 2027.