Isha Agarwal

Software Engineer

Bengaluru, Karnataka, India7 yrs 1 mo experience

Most Likely To SwitchAI Enabled

Key Highlights

4+ years of experience in compiler development for AI platforms.
Expertise in optimizing neural networks for edge devices.
Strong background in convolutional neural networks and image processing.

Stackforce AI infers this person is a specialist in AI-driven software development with a focus on compiler optimization.

Contact

hrishaaggarwal93@gmail.com LinkedIn

Skills

Core Skills

CompilerArtificial IntelligenceDeep Learning

Other Skills

Compiler OptimizationNeural NetworksImage Processing ModelsConvolutional Neural NetworksComputer VisionAutomation ScriptingConvolutional Neural Networks (CNN)Data StructuresC++AlgorithmsMachine LearningPython (Programming Language)Natural Language Processing (NLP)Artificial Intelligence (AI)Keras

About

Software Engineer with 4+ plus years of experience. Skilled in C++, data structures, and algorithms and have a strong knowledge of CNN. Worked on high-level compiler design for AI platforms. At Samsung R&D India Pvt. Ltd., I worked with compiler team for AI Edge devices to enable and optimize neural networks on NPU. Previous to Samsung, i worked at Intel India Pvt. Ltd., on optimizing compiler for Computer Vision accelerator, NNP-I 1000. Worked on compiler specially for (CNN) convolution neural networks for image processing models like ResNet, Dlv3, MobileNet etc. Used python to automate graph generation and testing. Post graduated from IIT Dhanbad. My focus is on producing clean, efficient code, optimizing software performance, and meeting project deadlines.

Experience

7 yrs 1 mo

Total Experience

1 yr 9 mos

Average Tenure

2 yrs 8 mos

Current Experience

Intel corporation

Compiler Engineer

Sep 2023 – Present · 2 yrs 8 mos · Bengaluru, Karnataka, India · On-site

LLVM Codegen compiler engineer

Samsung semiconductor

Associate Staff Engineer

Dec 2021 – Aug 2022 · 8 mos · Bengaluru, Karnataka, India

Worked on optimizing compiler for AI edge Devices.
Compiler is responsible for enabling and optimizing neural networks on NPU.
Worked on optimization for few of image processing models like Dlv3, Yolo, MobileNet, etc.
Handled L0-L1 Flush optimization for compiler and STM analysis for it.

Compiler OptimizationNeural NetworksImage Processing ModelsCompilerArtificial Intelligence

Intel corporation

2 roles

Deep Learning SW Development Engineer

Jul 2019 – Sep 2021 · 2 yrs 2 mos

Worked on compiler for Convolution Neural Networks for Computer Vision Hardware Accelerator Intel AI product ( NNPI-1 1000 )
Worked on different CNN (Standard CNN, DWC, and GCNN) for different data types int8 and fp16 to generate optimized solution to run on hardware efficiently.
Worked on ResNet50, ResNet101, ResNext, etc models.
Created shell and python scripts for automation and testing.
Used ONNX to generate graphs for testing.
Analyzed layer performance on hardware.
Handled post compiled flow, to read hardware blob, parse it and create graph with full data for driver to run.
Did performance analyzes for graphs.

Convolutional Neural NetworksComputer VisionAutomation ScriptingCompilerDeep Learning