Summary
Work History
Education
Skills
Websites
Timeline
Generic

Rishabh Srivastava

Bengaluru

Summary

Dynamic Senior Software Engineer with extensive experience at Vayavya Labs, specializing in autonomous driving simulations and perception pipelines. Proven track record of enhancing detection accuracy by over 12% using advanced technologies like TensorFlow and Deepstream. Strong analytical skills combined with effective communication foster collaboration and innovation in high-stakes environments.

Work History

Subcontractor L3

UPWORK
Remote
04.2024 - Current
  • End to End perception pipeline creation for entrypoint analytics.
  • Used NVIDIA Deepstream on Jetson platform.
  • Used Kafka, FastAPI and Flask servers for communications.
  • Effectively improved accuracy of detection by atleast 12% per detection.
  • Created stablisation pipelines for unstable RTSP and RTMP streams to reduce jitter and lags before feeding the same to Deepstream pipeline.
  • Created a Mail and Application based service which would send alerts to concerned entity by mail, whatsapp and in house android application with video of the alerts with the analytical information intact.
  • Created a Multi Camera RE-Identification ecosystem which could recognise people over multiple cameras using their vector embeddings.
  • Used Triton Inferencing server using docker for providing concurrent inference apis to remote services.
  • Use of GAN for increasing the accuracy of pixelated pictures in face detection and pipelines for enhancing the structure.
  • Used MLFlow for running model training scripts for production rollouts.
  • Technology Stack: Pytorch, Tensorflow and MLFlow for model training.
  • Deepstream C++ and Python for implementation.
  • OpenCV, Tensorflow for on device deployments using Python and C++.
  • FastAPI, Flask for analytics APIs.
  • Kafka and Redis for data communication and duplication avoidance.

Lead Engineer

VAYAVYA LABS
Benagaluru
02.2023 - 04.2024
  • Active part of Foretellix autonomous driving simulation project dealing with OpenScenario, OpenDrive and OpenPASS standards and language concepts.
  • Metric analysis for the important ADAS KPIs such as time to collision, present speed of collision etc.
  • Worked on Perception pipeline: Classification of Road Signs, Object Detection of Road Signs, cyclists, animals, big objects, stationary objects, other vehicles and Pedestrians etc.
  • Distance Estimation between Adjacent Cars for Tailgating. Lane Detection and Road Segmentation.
  • Using Vision Transformer to refine detection results.
  • Creation of augmented dataset and virtually generated data using TrafficSignGAN.
  • Creation of diverse situation traffic test cases using Stable Diffusion.
  • Created a competent and stable backend using Spring Boot for recording analytics and telemetry data.
  • Used Confluent Kafka pub-sub model and Redis for seamless telemetry communication.
  • Created NVIDIA Deepstream based pipelines for RTSP and RTMP source based inferencing.
  • Technology Stack: Pytorch and Various SOTA models.
  • OSC2 language for scenario generation.
  • OpenCV, Tensorflow for on device deployments using Python and C++.
  • Spring Boot for Telemetry data APIs.

Senior Software Development Engineer

LOCONAV
Gurugram
12.2021 - 02.2023
  • Company Overview: formerly Babtrucks
  • Inferencing Engine: Classification of Road Signs, Object Detection of Road Signs and Vehicles + Pedestrians.
  • Facial Expression Analysis, Distance Estimation between Adjacent Cars for Tailgating.
  • Lane Detection and Road Segmentation.
  • Drowsiness Detection for Driver Monitoring Systems.
  • Creation of traffic sign data using TrafficSignGAN.
  • Using DeepDrive for distraction monitoring.
  • Ability to run on Edge architecture having 1GB RAM and Quad Core CPU running on Armv7(Arm-V7A) and AArch64(Arm64-V8A) architectures on Android device.
  • Formerly Babtrucks
  • Technology Stack: YOLOV4 using TensorFlow Lite CPU, OpenCV DNN Darknet (JAVA & NDK C++).
  • SSD-MobileNet V2/V3 using TensorFlow Lite XNN PACK Delegate & TensorFlow Lite GPU Delegate. (KOTLIN + JAVA & C++ NDK).
  • Built Inferencing Modules for Deep Lab V3 based Road Segmentation using TensorFlow Lite GPU Delegate, TensorFlow Lite Interpreter + ARMNN Delegate. (JAVA + KOTLIN).
  • Achieved SSD MOBILENET V3 & YOLOV5 Object Detection using Tencent NCNN in GPU(VULKAN) and CPU. (KOTLIN & C++ NDK).
  • Currently working on Alibaba MNN Vulkan + CPU, Baidu PaddleLite GPU for Faster inferencing. (KOTLIN & C++ NDK).
  • Python Flask server based APIs for Driver monitoring system and ADAS data analytics.
  • Grpc and Kafka based communication module.
  • Non-Deep Learning Techniques such as HAAR Cascades.
  • Used C++ For Deployment, Python for Training in AWS EC2 instances, AWS S3 For Data Safekeeping.
  • Developed in QtCreator and Android Studio.

Senior Software Engineer

AGILIAD TECHNOLOGIES
Pune
06.2021 - 11.2021
  • Inferencing Engine: Constructed Multithreaded library in Visual Studio for Bone Model Inferencing using DICOM and TensorFlow.
  • Used C++ STL / Qt data structures to achieve Concurrent Inferencing for up to 500 MRI Slices.
  • UI and Widgets: Constructed C++ QWidgets IN Visual Studio using Docker in WSL2 and UI using XServer.
  • Widgets Responsible for 3D Look and Touch Sensitive UI for Displaying Bone Orthogonal View Pre-Surgery.
  • Custom QSS and CSS leveraged for each and every Widget for Industry Grade look.
  • Backend server using JDBC and Spring MVC.

Senior Software Engineer, Team Admin for AWS And Docker

QUALITAS TECH
Bengaluru
02.2020 - 02.2021
  • Cloud Module: Constructed Multithreaded library in generic C++ and QT for synchronization of Files between local machine / server and AWS S3 instances, AWS EC2 and Elastic Beanstalk.
  • Constructed a Java based application for data collection using android based edge devices.
  • Used STL / Boost/ Qt data structures to fine tune the system of synchronization and recover missing elements if any. Using AWS CPP SDK.
  • Camera Module: Constructed Multithreaded library in generic C++ and QT for Async Grabbing images based on software, line or continuous trigger using Machine Vision camera both monochrome and color.
  • Uses STL / Qt data structures to fine tune the system of GeniCam SDK and Implement Signal and Slots to receive data between various classes during the runtime.
  • Triggered multiple cameras connected to Edge Device such as NVIDIA Jetson Nano. Designed module to run in ARM/Linux/Windows Architecture.
  • Pre-Processing: Constructed a library for Cleaning, Resizing and Transformation of Image.
  • This module was designed by Using OpenCV and QImage Image Processing Libraries.
  • Inferencing Engine: Constructed a library in generic C++ and QT for a cross platform Deep learning inference Engine.
  • Engine designed using TensorFlow and Qt APIs. Engine is able to take in various frozen Neural Network models like FRCNN, ResNet50, Deeplab V2.0, Hardnet, SSD, YOLO v3 for inferencing the Images received using the Camera module in Edge device in ARM architecture.
  • Performed TensorFlow Inferencing in CUDA Based GPU. (CUDA 10.0, TensorFlow 1.15).
  • Performed Segmentation and Classification Concurrently on CPU.
  • Performed OpenVino Intel CPU optimized Segmentation, Classification and Object detection Concurrently. (QThread and QConcurrent).
  • Tested the Concurrent Module on AWS GPU instance. (G4DN and P3 instance Nvidia Tesla V100 and Intel Xeon 64 Core).
  • Analytics Engine: Created robust backend APIs using Django for Industrial analytics solutions and Spring Boot for Financial analytics solutions.
  • Docker Based Inferencing: Created a TensorFlow Docker / OpenVINO Docker for inferencing dataset using Ubuntu 18.04.

Intern

KREATIVPLM
Pune
03.2019 - 02.2020
  • Catia V5 Plugin Library Development.
  • Using OpenCV and Traditional PLM + Vision.
  • For Light System Design.
  • Using C++ and C libs for 3D imagery.

Education

Tech Diploma - PGDAC Program

CDAC
01.2018

B.E - Computer

MIT Pune
01.2017

Skills

  • C
  • Visual C#
  • Python
  • Java
  • JavaScript
  • HTML
  • CSS
  • CMake
  • QMake
  • MSVC
  • GCC
  • JRE
  • Python3
  • QT Creator
  • Visual Studio
  • Eclipse
  • Visual Studio Code
  • Android Studio
  • Ubuntu
  • Windows
  • WSL
  • Linux ARM
  • GitLab CI-CD
  • TortoiseSVN
  • MS-SQL
  • MySQL
  • QT
  • Kafka
  • Redis
  • Spring Boot
  • Spring Data
  • Vector Databases
  • Milvus
  • FastAPI
  • Django
  • Flask
  • OpenCV C Api
  • Deepstream C
  • TensorFlow C Api
  • Stable Diffusion
  • Huggingface
  • Deepstream Python
  • Triton Inference server
  • MLFlow
  • Darknet in OpenCV DNN
  • ONNX Runtime
  • TensorFlow Lite XNNPACK C Api
  • TFLite Android NDK C GPU Delegate
  • OpenCV Android NDK C Api
  • MediaPipe C Api
  • Tencent NCNN C Api
  • Nvidia TensorRT C Api
  • Intel OpenVINO C Api
  • QtCore
  • Shell
  • Batch
  • DockerFile
  • Maven
  • Anaconda
  • PIP
  • NuGet

Websites

Timeline

Subcontractor L3

UPWORK
04.2024 - Current

Lead Engineer

VAYAVYA LABS
02.2023 - 04.2024

Senior Software Development Engineer

LOCONAV
12.2021 - 02.2023

Senior Software Engineer

AGILIAD TECHNOLOGIES
06.2021 - 11.2021

Senior Software Engineer, Team Admin for AWS And Docker

QUALITAS TECH
02.2020 - 02.2021

Intern

KREATIVPLM
03.2019 - 02.2020

Tech Diploma - PGDAC Program

CDAC

B.E - Computer

MIT Pune

Similar Profiles

CREATE PROFILE
Rishabh Srivastava