Film · Research · AI Engineering

Nagaraj
Raparthi

Creative problem-solving has been the constant — the tools have just kept changing. I spent years modeling characters and environments for major Hollywood productions, translating a director's vision into technically precise geometry under real pressure. That discipline carried into GPU research published at SIGGRAPH, and now into building AI systems that generate and orchestrate visual content at MinIO. The visual instincts I developed in film are the same ones I bring to every system I design.

LinkedIn GitHub IMDb

Current Work

Software Engineer @ MinIO

November 2024 – Present

The same visual thinking that shaped film assets now shapes the AI pipelines I build. Creative work at a different layer of abstraction — but the eye for quality, the instinct for how things should feel, carries over directly.

Agentic AI Platform

Designed and built a multi-agent orchestration system on top of LLMs — handling complex analytical queries end-to-end with dynamic task decomposition, tool use, and structured output generation. Built on Go and Python with a custom agent runtime.

Visual AI Pipeline

Built an AI-driven content generation system that takes raw data and produces visual stories — imagery, comics, video. The same pipeline discipline I'd bring to a render farm, applied to vision-language models.

Full-Stack Infrastructure

Go backend with gRPC and REST, Python AI workers, React frontend with real-time visualization. Every layer built with the same attention to craft I learned in VFX production — nothing ships half-finished.

3
Concurrent AI systems in production Agentic pipeline, visual content engine, and real-time analytics frontend — all running simultaneously at MinIO

Research

Where I Learned to Bridge Art and Code

Move your cursor to guide the character
State IDLE
Vel 0.0
95%
Faster animation computation CPU-bound motion matching moved to GPU — SIGGRAPH Asia 2020
SIGGRAPH Asia 2020

GPU-Based Motion Matching for Crowds in Unreal Engine

At Texas A&M, I asked a question that came directly from my film experience: why does animating a crowd have to be so computationally expensive? Traditional motion matching is CPU-bound — sequential, unscalable. I moved the computation to the GPU and brought compute time down 95%, enabling real-time animation of hundreds of simultaneous characters. Published at SIGGRAPH Asia 2020. The bridge between artistic vision and systems performance — that's where I've always worked best.

View Paper Watch Demo

Background

The Journey

One throughline: finding where creative vision meets technical constraint, then building something at that exact intersection.

2024 – Present

MinIO

Software Engineer

  • Multi-agent AI orchestration platform
  • Visual content generation pipeline
  • Go · Python · React · gRPC

2022 – 2024

Apira Technologies

Unreal Engineer

  • Real-time metaverse avatar rendering
  • 2× frame rate improvement via GPU optimization
  • Artist-facing tool development

2020 – 2022

Val G. Hemming Simulation Center

Graphics Engineer

  • SIGGRAPH Asia 2020 — GPU-based motion matching
  • 95% reduction in animation compute time
  • W.A.V.E multi-projector alignment system

2018 – 2020

Texas A&M University

MS, Computer Graphics Technology

  • GPU optimization research
  • Motion matching thesis

2016 – 2018

Moving Picture Company (MPC)

3D Modeling Artist

  • Seven major film credits
  • Justice League · Ghost in the Shell · Detective Pikachu
  • Ad Astra · The Predator · Maleficent · Orient Express
8+
Bridging art and code Film production → GPU research → real-time graphics → agentic AI

Film Credits

Hollywood

Seven film credits as 3D modeling artist at Moving Picture Company (MPC). Every production taught me to make complex creative work hold up under scrutiny — at the pixel level, on a 60-foot screen.

Justice League

Justice League

2017

Motherboxes platform, Steppenwolf's helmet, climax structures, concept art

Ghost in the Shell

Ghost in the Shell

2017

Destroyed geisha suit, vehicles, 100+ optimized actor scans

Murder on the Orient Express

Murder on the Orient Express

2017

One of three modelers for the entire train

The Predator

The Predator

2018

Alpha predator (normal & armored), blend shapes topology, acting lead

Detective Pikachu

Detective Pikachu

2019

Sole modeler for the cheese-grater building

Ad Astra

Ad Astra

2019

Three spaceships (lead on one), helmets, space suit

Maleficent: Mistress of Evil

Maleficent: Mistress of Evil

2019

Aurora's crown and various hero assets

7
Major film credits at MPC 2016–2018 · Justice League, Ghost in the Shell, Detective Pikachu, Ad Astra, The Predator, Maleficent, Murder on the Orient Express
View on IMDb

Resume

Full professional background

A detailed view of my experience, skills, and publications across AI engineering, graphics research, and visual effects.

Download PDF

Contact

Let's work together

I'm at my best on problems that sit between disciplines — where visual craft meets systems engineering, where artistic intuition makes technical products better. If that sounds like what you're building, I'd love to talk.

Email nagarajraparthi31@gmail.com Phone +1 (682) 276-9799 LinkedIn nagaraj-raparthi GitHub TheOGSnakeKing