Prakhar Jagwani

Software Engineer

Atlanta, Georgia, United States0 mo experience

Key Highlights

  • Optimized ML model deployment at Samsung.
  • Boosted ML inference throughput by 30% at Windsurf.
  • Developed an innovative Android accessibility service at Microsoft.
Stackforce AI infers this person is a Backend-focused Software Engineer with expertise in Machine Learning and Mobile Development.

Contact

Skills

Core Skills

Machine LearningLarge Language Models (llm)Android Development

Other Skills

BashC++CUDAComputer NetworkingGitJavaJavaScriptLinuxONNXOpenCVParallel ProgrammingPyTorchPythonPython (Programming Language)React.js

Experience

Qualcomm

Software Engineer

May 2025Present · 10 mos · Santa Clara, California, United States · Hybrid

  • GPU Compiler - OpenCL Compute Performance

Georgia institute of technology

Graduate Teaching Assistant

Jan 2025Apr 2025 · 3 mos · Atlanta, Georgia, United States · On-site

  • Advanced Operating Systems

Windsurf

Software Engineer Intern

Aug 2024Dec 2024 · 4 mos · Mountain View, California, United States · On-site

  • Contributed to the ML inference backend:
  • Boosted throughput by 30% with 3X larger batch sizes, through dynamic capacity estimation and memory tracking
  • Reduced memory fragmentation by 50% with static and paged allocations; synchronized allocations across workers
  • Improved decode forward pass speed by 9% for a 70B model sharded 4 ways by adapting a custom AllReduce implementation (based on NVLink Sharp) to a distributed setting and integrating it into the inference server
  • Extended context length support to 128K tokens by optimally integrating an open source attention mechanism into the inference server
  • Implemented fine-tuning capabilities for a new LLM architecture by analyzing and adapting existing implementation
C++CUDAMachine LearningPyTorchLarge Language Models (LLM)

Samsung semiconductor

Software Engineer Intern

May 2024Aug 2024 · 3 mos · San Diego, California, United States · Hybrid

  • Contributed to the deployment and optimization of ML models on Exynos NPU:
  • Increased throughput by 7% on the Exynos NPU for an object detection model by modifying ML compiler code and optimizing execution schedules to improve prefetching
  • Assisted in the speech-to-text model selection process by assessing quality and performance across various ML configurations and comparing alternative approaches
  • Reviewed updates to the compiler for handling data transfer and frontend IRs, and presented detailed findings
C++PythonMachine Learning

Georgia institute of technology

Graduate Assistant

Jan 2024May 2024 · 4 mos · Atlanta, Georgia, United States · On-site

  • Part-time role at the College of Computing Data Centers:
  • Maintained observability for 600+ research and instruction servers by deploying and managing Sensu for monitoring
  • Automated system statistics reporting by creating Bash scripts that feed data into InfluxDB and Grafana dashboards
LinuxBash

Microsoft

Software Engineer Intern

Jun 2022Jul 2022 · 1 mo · Hyderabad, Telangana, India · Remote

  • Developed an Android accessibility service with 20+ customizable actions (inspired by iPhones Assistive Touch)
  • Supported foldable screens, landscape mode, lock screen functionality, animations, and extensive customization
JavaAndroid Development

Education

Georgia Institute of Technology

Master of Science - MS — Computer Science

Aug 2023May 2025

Indian Institute of Technology, Delhi

Bachelor of Technology - BTech — Computer Science

Jul 2019Jul 2023

Stackforce found 100+ more professionals with Machine Learning & Large Language Models (llm)

Explore similar profiles based on matching skills and experience