Gaurav Garg

CTO

India19 yrs 3 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in GPU computing and deep learning optimization.
Led multiple defense-related software projects.
Strong background in mentoring and team leadership.

Stackforce AI infers this person is a highly skilled software architect specializing in GPU computing and deep learning for defense and high-performance applications.

Contact

ugetgaurav@gmail.com LinkedIn

Skills

Core Skills

Deep LearningCudaHigh Performance ComputingGpu PerformanceComputer GraphicsGpu ArchitectureSoftware EngineeringTeam LeadershipSoftware Development

Other Skills

3D GraphicsASP.NETAlgorithmsApplication developmentArtificial Intelligence (AI)Bug fixesCC#C++CFDCUTLASSClient Requirements AnalysisClient project deliveryClient requirements mappingComputer Architecture

About

Technology Geek with passion for GPU Computing, Deep learning and 3D Graphics, Deep knowledge of Software Engineering & Architecture, Parallel Programming, GPU Architecture, Systems Architectures and low-level APIs (CUDA, OpenGL, Direct3D). Experience with System Software, Deep Learning, HPC applications, and 3D Visualization. Love to solve problems from first principles.

Experience

Nvidia

3 roles

Principal Software Architect, Deep Learning Inference

Promoted

Oct 2022 – Present · 3 yrs 5 mos

+ Deep learning inference performance optimization of some of the state-of-the-art generative AI models such as diffusion models, and LLMs
+ Writing optimized CUDA kernels for various deep-learning operators
+ Competitive performance analysis and making inference pipelines run faster on Nvidia GPUs
+ LLM inference optimization using techniques like kernel fusion, quantization, token speculation, etc.

Deep learning inference performance optimizationCUDAGenerative AILLM inference optimizationDeep Learning

Sr. HPC Developer Technology Engineer

Jun 2019 – Sep 2022 · 3 yrs 3 mos

+ GPU-accelerate and optimize applications in the domain of Molecular dynamics (GROMACS), CFD, Telecom, Cryptanalysis, and Weather models.
+ Experience with improving strong and weak scaling of applications on multi-GPU and multi-node systems.

GPU-accelerate applicationsMolecular dynamicsCFDTelecomCryptanalysisWeather models+2

Sr. Graphics System Software Engineer

Aug 2017 – May 2019 · 1 yr 9 mos

+ Work on DirectX user mode driver development which requires a detailed understanding of GPU architecture & ISA, Windows display driver model (WDDM), and Direct3D-based applications & engines.
+ Develop solutions that transparently scale application performance on multi-GPU systems. This involves performance analysis of applications, analyzing memory access patterns, and developing data and compute division strategies.

DirectX user mode driver developmentGPU architectureWDDMDirect3DComputer GraphicsGPU Architecture

Vizexperts

3 roles

Principal Software Architect

Promoted

Jul 2012 – Jul 2017 · 5 yrs

+ Responsible for product development, client project delivery, and supporting various sales & marketing teams on technology requirements
+ Hiring, training, and mentoring of the complete engineering team in the Bangalore center
+ Projects: Digital Sand Model, GeorbIS GIS, VizSim Engine, various Police, Army, DRDO & ISRO projects

Product developmentClient project deliveryMentoringSoftware EngineeringTeam Leadership

Lead Development Engineer

Jun 2010 – Jun 2012 · 2 yrs

+ INDIGIS 3D Project Development Lead
A GIS software targeted for the Army
Develop in-house expertise in the GIS domain
Adapt existing OSS tools for GIS usage and fill gaps, as necessary
Client requirements to features mapping and task scheduling
+ Advanced air defense Solution
Solution for demonstrating air defense simulations in distributed environment
User requirements (defense research labs) to implementation and deployment
+ vSite3D Development Lead
Solution for Remote visualization and collaboration specially for O&G 3D applications
Complete technical ownership
Redesigning and re-implementing major parts of software as per changing requirements (Windows support, scalability, performance)