Avinash Paliwal

Co-Founder

Seattle, Washington, United States8 yrs 5 mos experience

Highly StableAI Enabled

Key Highlights

Expert in 3D Computer Vision and Generative Video Models.
Published in top-tier conferences like CVPR and SIGGRAPH Asia.
Led innovative projects in video generation and scene reconstruction.

Stackforce AI infers this person is a leading expert in AI-driven video generation and 3D computer vision.

Contact

Skills

Core Skills

Video Generation3d Computer VisionMachine LearningDeep LearningComputer VisionResearchData AnalysisEmbedded Systems

Other Skills

PythonDiffusion Models3D Gaussian SplattingNovel View SynthesisReal-Time SystemsStereo View-Time Interpolation3D DisplaysJavaSQLRTOS DevelopmentDebuggingGenerative AICMatlabHTML

About

Founding AI Researcher at Morphic Inc., working at the intersection of 3D Computer Vision and Generative Video Models. I lead the development of production-grade video generation features, such as image-to-video (3D Motion) and video-to-video camera control. My expertise centers on diffusion models and scene reconstruction/generation using techniques like 3D Gaussian Splatting. I hold a Ph.D. from Texas A&M University , with prior research experience at Meta Reality Labs and Leia Inc.. My work has been published in top-tier conferences including CVPR, ICCV, ECCV, and SIGGRAPH Asia.

Experience

Nuance labs

Founding Research Scientist

Mar 2026 – Present · 0 mo · Seattle, Washington, United States

Morphic

Founding AI Researcher

Apr 2025 – Mar 2026 · 11 mos · San Jose, California, United States

Led development of production video generation features, including 3D Motion for custom camera videos from a single image, and video inpainting/outpainting.
Building an upcoming feature based on “Reshoot-Anything” (under review): a self-supervised video-to-video camera control model trained on arbitrary monocular videos without 3D annotations or static/dynamic scene filtering.
Trained large video models using multi-node distributed training with multiple fine-tuning strategies, including full fine-tuning, LoRA, and context adapters, enabling faster experimentation and improved controllability.
Built parallelized data processing pipelines for video curation, filtering, and preprocessing to support robust training and production workflows.
Exploring agentic text-to-trajectory planning with LLMs/MLLMs to generate scene-aware camera paths from prompts, with optional human-in-the-loop refinement via intermediate visualizations and text/3D edits.

Video Generation3D Computer VisionMachine LearningDeep LearningPython

Leia inc.

Research Intern

Sep 2021 – Dec 2021 · 3 mos · College Station, Texas, United States · Remote

Developed a near real-time stereo view-time interpolation method for handheld 3D displays using multi-plane disparities and non-uniform coordinates (CVPR 2023).

3D Gaussian SplattingReal-Time SystemsComputer Vision

Texas a&m university

3 roles

Graduate Teaching Assistant

Aug 2020 – Dec 2020 · 4 mos · College Station, Texas, United States

CSCE 441 Computer Graphics

Graduate Teaching Assistant

Jan 2020 – May 2020 · 4 mos · College Station, Texas, United States

CSCE 489/689 Computational Photography

Graduate Research Assistant

Sep 2019 – Mar 2025 · 5 yrs 6 mos · College Station, Texas, United States

Developed a single-image 360° 3D scene generation pipeline using diffusion-based panorama/depth synthesis followed by 3D Gaussian Splatting optimization (SIGGRAPH Asia 2025).
Proposed a modular pixel reshading network that predicts view-dependent shading (e.g., moving specular highlights) for novel views, improving single-image novel view synthesis realism (SIGGRAPH Asia 2023).
Published additional work on dynamic scene interpolation (WACV 2023), GAN-based raw video denoising (ICCP 2021), and hybrid imaging slow-motion reconstruction (TPAMI 2020).

Diffusion Models3D Gaussian SplattingComputer VisionResearch

Amazon

Software Development Engineer Intern

May 2019 – Aug 2019 · 3 mos · Seattle, Washington

Built a Java/SQL pipeline to analyze customer interaction history and generate personalized recommendations to improve user engagement.

Stereo View-Time Interpolation3D DisplaysComputer Vision

Bajaj auto ltd

Senior R&D Engineer

Jul 2016 – Jul 2018 · 2 yrs · Pune/Pimpri-Chinchwad Area

Developed core modules for BALOS (company’s first in-house RTOS), including OTA bootloader, fault management, and a lightweight CAN protocol reducing bandwidth by 40%.
Built an emulation and debugging environment to run the application + RTOS stack without hardware, enabling faster root-cause analysis and iteration.

JavaSQLData Analysis