Rohit Saxena

Senior Software Engineer

Bengaluru, Karnataka, India7 yrs experience

Most Likely To SwitchHighly Stable

Key Highlights

Expert in deep learning model deployment and optimization.
Proven track record in computer vision applications.
Strong background in algorithm design and analysis.

Stackforce AI infers this person is a Deep Learning and Computer Vision specialist focused on Embedded Systems.

Contact

saxenarrohit@gmail.com LinkedIn

Skills

Core Skills

Deep LearningComputer VisionModel OptimizationResearch

Other Skills

Deep Learning Model DeploymentDNN QuantizationConvolutional Architecture DevelopmentImage ClassificationObject DetectionImage SegmentationDepth EstimationC++PythonPytorchTfliteQuantizationLiterature ReviewNumber Theory ResearchFront-end Web development

About

Experienced Software Engineer with a demonstrated history of working in Deep Learning Deployment softwares and Deep Learning Application in Embedded Systems. Around 1.5 years of work experience in immensely popular Deep Learning Model Compression technique called Quantization, comprising important Research and Development contributions leading to key KPI improvements. Current work is involved in deep learning based Computer Vision application, primarily focussing on Obstacle Detection task, targeting Embedded systems. With a strong grip on Design and Analysis of Algorithms, high proficiency in languages like C++, Python and Pytorch along with a sound knowledge of Neural Networks and Deep Learning, I am looking forward to create Computer Vision applications with great commercial value. Fields of interest:- Software Development, Deep Learning, Computer Vision, Data Structures and Algorithms, Data Science Work Experience:- Deep Learning Model Deployment, DNN Quantization, Convolutional Architecture Development, Image Classification, Object Detection, Image Segmentation, Depth Estimation Languages:- C++, Python Deep learning Libraries:- Pytorch, Tensorflow, Tflite Deep Learning Deployment SDK:- Armnn(https://github.com/ARM-software/armnn), ComputeLibrary(https://github.com/ARM-software/ComputeLibrary), MKL-DNN(https://github.com/rsdubtso/mkl-dnn), Tensorflow(https://github.com/tensorflow/tensorflow)

Experience

7 yrs

Total Experience

2 yrs 4 mos

Average Tenure

5 yrs 4 mos

Current Experience

Samsung r&d institute india - bangalore private limited

Senior Software Engineer

Jan 2021 – Present · 5 yrs 4 mos · Bengaluru, Karnataka, India

Constructed light-weight Regression and Classification Convolutional architecture and implemented deep learning based Obstacle detection on highly resource constrained Micro Controller hardware unit.
Achieved a very high binary classification accuracy over a huge target dataset within a very low memory and computation budget associated with the target hardware.
Developed Pytorch based modules for training, validation, data pipelines, visualization and prediction pattern analysis of Obstacle detection model.
Developed Tflite based Obstacle Detection C++ application to run the Obstacle detection model on the target Micro-Controller hardware.
Extensively researched the state of the art Image Classification, Segmentation and Depth estimation architectures, covered all the important research papers in field of Obstacle detection and classification.

Deep Learning Model DeploymentDNN QuantizationConvolutional Architecture DevelopmentImage ClassificationObject DetectionImage Segmentation+7

Samsung r&d institute india

Machine Learning Engineer

Aug 2019 – Jan 2021 · 1 yr 5 mos · Bengaluru, Karnataka, India

Enhanced indegenous C++ based quantization modules by implementing innovative linear quantization schemes, improved the accuracy of the quantized models compared to the conventional quantization schemes.
Fixed runtime bugs associated with the quantization of the high precision(32-bit) deep learning model and low precision(8-bit) quantized model inference.
Extensive literature review of the state of the art Deep learning Quantization techniques covering all the important research papers and an in-depth analysis of the corporate research in the field.

C++QuantizationDeep LearningLiterature ReviewModel Optimization