Arnav Agharwal

Senior Software Engineer

San Francisco, California, United States14 yrs 3 mos experience

Highly StableAI ML Practitioner

Key Highlights

8+ years in AI/ML system development.
Led initiatives at Google and NVIDIA.
Expert in performance optimization and healthcare AI.

Stackforce AI infers this person is a highly skilled AI/ML engineer with extensive experience in healthcare technology.

Contact

Skills

Core Skills

Machine LearningArtificial Intelligence (ai)

Other Skills

C++Python (Programming Language)Docker ProductsGoogle Cloud Platform (GCP)PyTorchTensorFlowDICOMData GovernanceTerraformApache BeamLinuxNumPyGitGoogle Cloud DataflowTechnical Computing

About

Senior Software Engineer with 8+ years building production AI/ML systems from research to scale. Led technical initiatives in foundation model inference, multimodal AI platforms, and distributed ML infrastructure at Google Research and NVIDIA. Proven track record delivering measurable impact in healthcare AI, LLM/VLM serving, and high-performance inference optimization.

Experience

14 yrs 3 mos

Total Experience

3 yrs 2 mos

Average Tenure

1 yr 6 mos

Current Experience

Nvidia

Senior Software Engineer

Dec 2024 – Present · 1 yr 6 mos

Tech Lead – Distributed Multimodal Processing: Enabled robust, scalable (data sharding), distributed (Cloud-ready) AutoML-style intelligent hyperparameter search by re-architecting monolithic software (PipeTuner), accelerating multimodal inference pipeline performance tuning for product teams and enterprise customers.
Tech Lead – Foundation Model Infrastructure: Modernized the software architecture of NVIDIA Inference Microservice (NIM) for LLMs/VLMs, accelerating deployment of next-gen GPU architectures (incl. Blackwell) and reducing go-to-market cycles from months →weeks.
Production AI Systems: Productized NVILA models to a vLLM-based backend, delivering substantial inference throughput acceleration over research and seamless Cloud integration.
Performance Optimization: Improved GPU ROI by 10x video processing throughput using systematic bottleneck analysis and early adoption of the NVIDIA’s Video Codec SDK.

C++Python (Programming Language)Docker ProductsGoogle Cloud Platform (GCP)PyTorchTensorFlow+2

Google

2 roles

Senior Software Engineer

May 2022 – Sep 2024 · 2 yrs 4 mos · San Francisco Bay Area

C++Python (Programming Language)DICOMData GovernanceTerraformApache Beam+9

Software Engineer

Apr 2019 – May 2022 · 3 yrs 1 mo · San Francisco Bay Area

Google Health (https://health.google)

C++Python (Programming Language)DICOMData GovernanceTerraformApache Beam+7

Heartflow, inc

2 roles

Imaging Algorithms Engineer

Oct 2017 – Mar 2019 · 1 yr 5 mos

C++Python (Programming Language)DICOMLinuxGitPyTorch+2

Software Engineer

Oct 2017 – Mar 2019 · 1 yr 5 mos

Production ML Systems: Developed production ML pipelines for cardiac CT scans, enabling major product release with improved diagnostic accuracy.
DevOps Innovation: Designed a bespoke CI/CD framework, reducing regressions and accelerating deployment cycles.