Vedavyas Potnuru

Software Engineer

Seattle, Washington, United States7 yrs 6 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in machine learning and deep learning technologies.
Proven track record in optimizing AI models for real-world applications.
Strong background in software engineering and hardware design.

Stackforce AI infers this person is a Machine Learning Engineer with expertise in AI model optimization and hardware integration.

Contact

potnuru.vedavyas@gmail.com LinkedIn

Skills

Core Skills

Machine LearningSoftware EngineeringAugmented RealityDeep LearningMedical ImagingElectrical EngineeringHardware Design

Other Skills

3D CNN modelARM Compute LibraryAlgorithmsAutoCADCC (Programming Language)C++CaffeDeep Generative Adversarial NetworkElectronicsImage ProcessingJavaML/AI inferenceMesh TensorFlowMicrosoft Copilot

Experience

Microsoft

Senior Software Engineer (Microsoft Turing)

Oct 2021 – Present · 4 yrs 5 mos · Seattle, Washington, United States

Microsoft Copilot
LLM/GPT stuff, ML/AI inference and optimization

Microsoft CopilotML/AI inferenceoptimizationMachine LearningSoftware Engineering

Snapchat

Machine Learning Intern - Augmented Reality

Apr 2021 – Jun 2021 · 2 mos · Los Angeles, California, United States

Working on image-to-image Deep Generative Adversarial Network (GAN) model compression techniques to create SnapML templates that help users create mobile friendly lenses

Deep Generative Adversarial NetworkSnapML templatesMachine LearningAugmented Reality

Intel corporation

Graduate SW Intern - Deep Learning

Jun 2020 – Sep 2020 · 3 mos · San Francisco Bay Area

Worked on large 3D CNN model training and inference for medical image segmentation (leveraging Mesh TensorFlow)

3D CNN modelmedical image segmentationMesh TensorFlowDeep LearningSoftware Engineering

Samsung r&d

Software Engineer

Aug 2016 – Sep 2019 · 3 yrs 1 mo · India

Areas : Software Development, On Device AI, Deep Learning, Open source DL Frameworks
C, C++, Python
Contribute to TensorFlow-Lite source
8-bit fixed point quantization of Neural Network models
Quantization aware training(QAT) and post training quantization(PTQ)
Evaluate and improve third party NPU SDK’s Quantization feature
Application to detect objects in videos using Neural Networks(CNNs) on Smart TV platform
Optimize the inference using ARM Compute Library (C++) to accelerate the application by 6x targeting ARM Mali GPU.
Contribute to ARM compute library source and fix bugs
Bring Caffe framework source to use and perform memory and compute optimization of different NN layers for the smart TV embedded platform.

Software DevelopmentOn Device AIDeep LearningTensorFlow-LiteQuantizationSoftware Engineering

Texas instruments

Summer Internship

May 2015 – Jul 2015 · 2 mos · Bengaluru Area, India

Designed an RTL model for interfacing with the SAR ADC correlating the specification of Suvega device developed by Texas Instruments
Created models of SAR ADCs which differ in their Analog Input modes implementing the Successive Approximation Register logic and eventually verified the working logic of the simulated design

RTL model designSAR ADCSuccessive Approximation Register logicElectrical EngineeringHardware Design