satrajit@rutgers
.----. ────────────────────────────────────────────────
.---------. | == | OS: MS ECE (Machine Learning)
|.-"""""-.| |----| University: Rutgers University - New Brunswick
|| || | == | GPA: 4.0 / 4.0
|| || |----| Research: Graphics | GPU | Neural Rendering
|'-.....-'| |::::| Exchange: Princeton University
`"")---(""` |___.| ────────────────────────────────────────────────
/:::::::::::\" _ "
/:::=======:::\`\`\ Languages.Systems: C++20, C, CUDA C++
jgs `"""""""""""""` '-' Languages.ML: Python, PyTorch, TensorFlow
Languages.Web: JavaScript, TypeScript
Languages.Shading: GLSL, HLSL
────────────────────────────────────────────────
Graphics.APIs: OpenGL 4.6, Vulkan, CUDA 12.x
Graphics.Techniques: Render Graphs, PBR
Graphics.Research: GPU Scheduling, Interop
────────────────────────────────────────────────
ML.Vision: NeRF, 3D Gaussian Splatting
ML.Domains: Scene Reconstruction, Neural Render
ML.Multimodal: Audiovisual Perception
────────────────────────────────────────────────
Tools.Build: CMake, Ninja, Git
Tools.Debug: RenderDoc, NVIDIA Nsight
Tools.Systems: Linux, WSL2, Docker
────────────────────────────────────────────────
┌─────────────────────────────────────────────────────────────────────┐
│ │
│ Real-Time Rendering Systems │
│ └─ GPU scheduling algorithms and render graph optimization │
│ └─ Hardware-software co-design for graphics pipelines │
│ └─ CPU-GPU parallelism and synchronization primitives │
│ │
│ Neural Rendering │
│ └─ Neural radiance fields and 3D Gaussian splatting │
│ └─ Differentiable rendering and inverse graphics │
│ └─ Efficient training for real-time applications │
│ │
│ Computational Photography │
│ └─ Physics-based image relighting and material capture │
│ └─ Multi-view geometry and camera calibration │
│ └─ Structure from motion and dense reconstruction │
│ │
│ Immersive Systems │
│ └─ VR/AR rendering pipelines and optimization │
│ └─ Real-time physics simulation on GPU │
│ └─ Human-computer interaction in virtual environments │
│ │
└─────────────────────────────────────────────────────────────────────┘
┌──────────────────────────────────────────────────────────────────────┐
│ PROJECT : Custom C++ Game Engine & GPU Scheduling Research │
│ STATUS : Active Development │
│ FOCUS : Performance Measurement & Optimization │
├──────────────────────────────────────────────────────────────────────┤
│ │
│ Architecture │
│ ├─ DAG-based render graph with explicit pass dependencies │
│ ├─ Automatic resource lifetime tracking and barrier insertion │
│ ├─ Multithreaded work-stealing task scheduler │
│ └─ Lock-free job queue for CPU-GPU overlap │
│ │
│ GPU Interoperability │
│ ├─ CUDA-OpenGL zero-copy buffer sharing │
│ ├─ Direct GPU buffer writes via mapped memory │
│ └─ Synchronized texture updates between compute and graphics │
│ │
│ Instrumentation │
│ ├─ Per-frame timing with CPU and GPU timestamps │
│ ├─ Command queue depth and occupancy metrics │
│ ├─ Synchronization overhead profiling │
│ └─ GPU utilization and memory bandwidth analysis │
│ │
│ Goal: Identify scheduling bottlenecks through empirical measurement │
│ │
└──────────────────────────────────────────────────────────────────────┘
┌──────────────────────────────────────────────────────────────────────┐
│ PROGRAM : Princeton University Research Exchange │
│ DURATION : 2024 │
│ FOCUS : Neural Rendering Techniques │
├──────────────────────────────────────────────────────────────────────┤
│ │
│ Neural Radiance Fields (NeRF) │
│ └─ Volumetric scene representation with neural networks │
│ └─ Differentiable volume rendering for novel view synthesis │
│ └─ Training strategies for sparse and dense view scenarios │
│ │
│ 3D Gaussian Splatting │
│ └─ Explicit 3D scene representation with Gaussian primitives │
│ └─ Real-time rendering through rasterization │
│ └─ Optimization for high-quality reconstruction │
│ │
│ Neural Character Generation │
│ └─ Animatable 3D avatar creation from limited data │
│ └─ Neural skinning and deformation models │
│ └─ Real-time rendering of dynamic human characters │
│ │
└──────────────────────────────────────────────────────────────────────┘
┌────────────────────────────────────────────────────────────────────┐
│ │
│ VR/AR Systems Portfolio │
│ ├─ VR Fitness Application │
│ │ └─ 3rd Place, Google x GeeksforGeeks "Solving for India" │
│ ├─ VR Warehouse Training System │
│ │ └─ Interactive learning modules and spatial annotation │
│ ├─ Architectural Visualization Platform │
│ │ └─ Real-time walkthroughs with dynamic lighting │
│ └─ Virtual Cinema Experience │
│ └─ Social VR environment with synchronized playback │
│ │
│ Computer Vision Projects │
│ ├─ Geometry-Guided Image Relighting │
│ │ └─ Extended StyLitGAN with physics-based constraints │
│ ├─ Stereo Reconstruction System │
│ │ └─ Disparity estimation and 3D point cloud generation │
│ └─ Multi-View Geometry Pipeline │
│ └─ Epipolar geometry and structure from motion │
│ │
│ Software Engineering │
│ ├─ Code & Conquer - Gamified Coding Platform │
│ │ └─ Full-stack with Docker execution engine, Stripe payments │
│ ├─ Quantitative Trading Infrastructure │
│ │ └─ Real-time data, backtesting, workflow automation │
│ └─ Physics Simulation with CUDA │
│ └─ Parallel rigid body dynamics and collision detection │
│ │
└────────────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────────┐
│ │
│ Master of Science in Electrical & Computer Engineering │
│ Rutgers University - New Brunswick │
│ Expected: May 2025 | GPA: 4.0 / 4.0 │
│ │
│ Concentration: Machine Learning │
│ │
│ Relevant Coursework: │
│ ├─ Machine Vision │
│ ├─ Software Engineering │
│ ├─ Programming for Finance │
│ ├─ Computer Graphics │
│ ├─ GPU Computing │
│ └─ Deep Learning │
│ │
│ Research Exchange: │
│ └─ Princeton University - Neural Rendering & 3D Reconstruction │
│ │
└─────────────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────────┐
│ │
│ Email : sg2231@rutgers.edu │
│ LinkedIn : linkedin.com/in/satrajit-ghosh │
│ GitHub : github.com/satrajitghosh183 │
│ │
└─────────────────────────────────────────────────────────────────────┘
Last Updated: November 2024