Rome, Italy Robotics · 3D Vision

3D Vision & Robotics Researcher

Ph.D. · Former Postdoc · Industry Researcher

I build end-to-end autonomous systems at the intersection of SLAM, active 3D perception, and robotic platforms (UAVs and mobile manipulation). Previously at the Italian Institute of Technology (ADVR & CCHT), and currently in industry developing drone-based 3D vision systems.

Download CV GitHub Google Scholar LinkedIn Contact

Current

Autonomous drone-based 3D vision and SLAM pipelines for agricultural field analysis.

Focus

SLAM · 3D mapping · neural scene representations · multimodal fusion · real-time autonomous systems.

Highlights

1st prize: VISUM 2021 & EEML 2020 summer school.

What I Build

Systems that connect perception, planning, and deployment — not just standalone models.

Autonomous Robots

UAVs and robot manipulators that execute missions end-to-end: sensing, decision-making, and safety-aware operation.

PinocchioPX4MAVSDK/MAVLinkROS 2

Active 3D Perception

SLAM and 3D mapping pipelines (including neural scene representations) from small to large-scale, real-world environments.

SLAM3D Gaussian SplattingNeRF

End-to-End Engineering

Data logging, visualization, reproducible experiments, and deployment constraints (real-time, safety, robustness, maintainability).

Python/C++ViserGPU/Edge pipelines

Flagship Systems & Projects

A curated selection of systems I’ve designed and developed. Visuals are included where available, and placeholders indicate future drop-in demos.

[Placeholder: field mapping / trajectory / reconstruction]

Drone-based 3D Vision for Agricultural Field Analysis

Industry · Rome (Current)

Developing SLAM and 3D reconstruction pipelines for large-scale field scanning, combining autonomous missions, real-time RGB streaming, and benchmarking workflows.

PX4MAVSDKVOXL 2SLAM3D Reconstruction

Real-time SLAM with 3D Gaussian Splatting (RIAP)

Postdoc · ADVR, IIT Genoa

Built real-time SLAM pipelines based on neural scene representations to support immersive teleoperation and mapping through VR.

Neural SLAM3DGSReal-timeGPU

Autonomous Robotic 3D Artifact Scanning (CTE-Genova)

Postdoc · CCHT, IIT Venice → Deployed at CTE-Genova

Project

Developed an autonomous scanning pipeline using UR manipulators and 3D sensing, including 6-DoF pose tracking for complex artifacts. System delivered and deployed on-site.

UR5e/UR3e6-DoF Tracking3D ScanningAutomation

[Placeholder: multimodal fusion diagram + qualitative detection]

Multimodal 3D Scene Perception (PhD Work)

PAVIS, IIT · Thesis: Multimodal 3D Scene Perception

Research on LiDAR–camera fusion and multi-view 3D understanding for robust scene perception and object localization.

LiDAR–RGB fusion3D DetectionMulti-viewLarge-scale

Startup Direction: Autonomous Product Digitization

I’m prototyping an autonomous robotic platform that digitizes physical products into standardized, high-quality 3D assets with minimal supervision.

The Core Idea

Product digitization today is often manual, expensive, and hard to scale. The key insight is to treat digitization as an embodied intelligence problem: the robot actively decides where to move and what to observe, adapting to each object and scene.

System (high level)

Mobile robotic platform with integrated perception and actuation
Localization and scene understanding for object-aware interaction
Adaptive data acquisition driven by scene geometry
Automated generation of standardized 3D assets

This section outlines a research-driven product direction; technical details and system design are intentionally abstracted at this stage.

[Placeholder: architecture diagram / concept illustration]

Target Applications

E-commerce catalogs Product visualization XR/AR content Marketing assets

Collaboration

I’m exploring funding and collaboration opportunities (technical + business). If you’re working on digitization pipelines, 3D commerce, or robotics productization, feel free to reach out.

Publications

Peer-reviewed and preprint work across 3D perception, multimodal fusion, and robotics.

From Fields to Splats: A Cross-Domain Survey of Real-Time Neural Scene Representations

arXiv 2025 · Javed Ahmad, et al.

[arXiv]

Hands-Free Heritage: Automated 3D Scanning for Cultural Heritage Digitization

arXiv 2025 · Javed Ahmad, et al.

[arXiv]

AAPOE: Automated Artifacts Position and Orientation Estimation in Cultural Heritage

IEEE/ASME MESA 2024 · Javed Ahmad, et al.

[Paper] [Teaser] [Presentation]

mmFUSION: Multi-modal Fusion for 3D Object Detection

arXiv 2023 · Javed Ahmad, Alessio Del Bue

[Paper] [Presentation] Code available upon request.

Fill in the Blank for Fashion Complementary Outfit Product Retrieval

Machine Vision and Applications (2023) · Author list incl. Javed Ahmad

[Paper] [Code]

Multi-view 3D Object Localization from Street-Level Scenes

ICIAP 2022 · Javed Ahmad, Matteo Toso, Matteo Taiana, Stuart James, Alessio Del Bue

[Paper] [Code] [Presentation]

Background

A quick snapshot of my trajectory across research and industry.

Experience

Computer Vision Researcher · DeepPlants (Rome)

Sep 2025 – Present

SLAM + 3D reconstruction pipelines for agricultural field analysis; autonomous drone missions and data workflows.

Postdoctoral Researcher · IIT (ADVR & CCHT)

Feb 2024 – Aug 2025

3D Gaussian Splatting SLAM; autonomous robotic scanning for cultural heritage; 6-DoF tracking.

PhD Researcher · IIT & UniGe (PAVIS)

Nov 2020 – Dec 2023

Multimodal 3D scene perception and multi-view 3D localization.

Skills

Core

PythonC++PyTorchOpenCVOpen3DViser

Robotics

ROS 2MoveIt 2PinocchioUR-RTDE PX4MAVSDKGazebo SITL

3D / Mapping

SLAMCOLMAPNeRF3D Gaussian Splatting MMDetection3D