Every THING is Internet of Thing

Projects

Projects

Edge IoT Framework primarily for Life Sciences Use-cases:

  • Role: Solutions Architect (3 Months)
  • Skills: Service Oriented Architecture, Azure IoT Edge, AWS IoT Greengrass, Docker, Kubernetes
  • Accomplishments:
    • Currently evaluating AWS Outpost, AWS EKS, Kubernetes, Containerized deployments, Apache Kafka, Pravega, Scalable MQTT brokers deployment to address high volume sensors, high-definition video data use-cases in 5G Edge Computing scenarios.
    • Worked on common Edge IoT use-cases, various possible scenarios considering data flows, data types, data restrictions, privacy, latency, bandwidth consumptions, connectivity restrictions, etc., primarily revolving around Life Sciences projects and devices.
    • Evaluated Open-Source Edge Projects such as KubeEdge, ioFog, EdgeX, LF-Edge Umbrella projects against identified use-cases.
    • Evaluated suitability of AWS IoT Greengrass and Azure IoT Edge + other AWS/Azure on-premises services for various Edge computing scenarios, presented pros & cons of both public cloud platforms and created various possible use-cases architecture/design with AWS/Azure as primary components of framework.
    • Architecting a common framework based on open-source Edge projects which can complement the public cloud services in on-premises Edge computing scenarios.

Jetson Nano based Healthcare IoT Device as a Guided Pipetting Tip Sensing System:

  • Role: Systems Engineer (6 Months)
  • Accomplishments:
    • Carried out PoC tasks like flashing boards to simulate mass flashing at factory.
    • Multiple PoCs to understand customizing RootFS, Secure Boot, Bootloader Splash Screen.
    • Interfacing Bluetooth module with NVIDIA L4T BSP software for Jetson Nano. All these PoC tasks resulted in a concrete plan to be executed at factory manufacturing.
    • Evaluated, Designed and implemented Firmware update and OS update mechanism based on Mender-Yocto Open-Source project.
    • Designed and implemented device side C++ & Python, RESTful HTTP protocol-based multiprocessor-distributed IoT connectivity application for features like device identity & registration, status, *user-device association*, *certificates management*, *device shadow*.
    • Evaluated different inter-process communication tools, RPC mechanisms as ROS1, ROS2, D-Bus, ZMQ to Architect & Design multiprocessor-distributed application.

STM32 MCU based Portable COVID-19 Diagnostic device kit:

OTA Firmware Updates for a STM32 MCUs and full OS Updates for x86 carrier boards:

  • Role: IoT Developer (14 Months)
  • Accomplishments:
    • Designed and implemented custom bootloader with Dual bank strategy for STM32F407 MCU to achieve robust Firmware update requirements. MAVLink communication protocol and Signature & checksum verification were few of the other key features implemented.
    • Designed & developed C++ & Python based multithreaded multiprocessor-distributed application to achieve OTA firmware update for multiple STM32 MCUs. AWS IoT Jobs, Device Shadow, Secure communication and UART based serial communication were key features.
    • PoC for full OS image and Application update using Mender.io Open-Sorce project. Full OS OTA updates with Mender server hosted on AWS EC2. Also, same kind of updates with USB and over LAN were also achieved with local Python server.
    • Evaluated and finalized Ayla IoT platform for early market release without full fledge cloud development. Device provisioning, status, firmware updates to multiple devices, sensor data streaming were key features achieved in 4 months duration.
    • Designed and implemented a NodeJS module for Data Synchronization between device and cloud using MongoDB Realm and MongoDB Atlas cloud databases. Understanding of new platform and successful delivery was achieved in 3 months.

    • Multiple product devices are deployed in the field Product Demo.

Drone-based Asset Inspection with AWS IoT Greengrass & AWS Robomaker Services:

  • Role: Robotics, IoT Developer (8 Months)
  • Accomplishments:
    • Understanding of the new to be launched or recently launched AWS Services like Robomaker, Greengrass, AWS IoT and their use-cases for Robotics projects were achieved in 3 months with demonstratable applications as an outcome.
    • Robotics framework ROS based distributed application had Machine Learning features like Rust and leakage detection models that were deployed with AWS IoT Greengrass to multiple devices as NVIDIA Jetson TX2 (drone), mobile robots and x86 machines.
    • AWS IoT device shadow updates and IoT Jobs for firmware update were also used.
    • AWS Lambda functions were deployed to AWS IoT Greengrass to run ML inference.
    • Robotics application was able to capture and upload the thermal and normal camera videos to AWS S3 buckets, which were consumed by AWS Sagemaker for Machine Learning training.
    • NodeJS, JavaScript based Web Application running on AWS EC2 was developed to control Robots with commands, to trigger firmware updates, and to initiate ML training in Sagemaker.
    • Camera Live video streams were transmitted to AWS Kinesis Video Streams and the same were rendered on Web Application in HLS format.
    • Drone simulation around oil-rig running in background in AWS official Tech Talk

    • Demos were successfully showcased at CERAWeek 2019 and AWS re:MARS 2019 events

Robotics Device Management Effort Estimation & Proposal Creation:

  • Role: Developer (2 Months)
  • Responsibilities:
    • Understand the Product Requirements
    • Understand various Azure IoT & other Cloud services and mapping them to Product Requirements.
    • Create an Overview ROS & Azure Cloud based Architecture using Docker Container.
    • Create Docker Containerized ROS Demo applications running on x86 Linux machine accessing connected peripheral devices, file system.
    • Create complexity sheet with various modular tasks to arrive at an effort estimation.
    • Coordinate with other Product teams involved.

Robotics ROS Framework Packages Migration:

  • Role: Developer (6 Months)
  • Accomplishments:
    • Understand the Open-Source Robotics frameworks ROS1 & ROS2.
    • Understand Robotics & Cloud Packages with respect to original architecture and implementations of assigned package.
    • Create package architecture with respect to new ROS2 framework, Migrate packages from ROS1 framework to ROS2 framework.
    • Follow up & Interact with Open-Source community, accommodate community suggestion.
    • Get the migration work Reviewed & Accepted by the Community members.
    • Successfully migrated vision-opencv ROS package.

IBM Content Collector (part of IBM Enterprise Content Management):

  • Role: Module Lead Engineer (18 Months)
  • Responsibilities:
    • Module Lead Engineer for IBM Content Collector for Email (primarily for IBM Notes/Domino).
    • Developed Domino & Notes Application for Content Collector Expiration Manager module using Lotus Scripts.
    • Debugged and resolved multiple server-side crashes, which improved stability of the product. resulting in retaining the customer to continue use IBM Domino/Notes applications.
    • Solved crucial Application Recovery issue with Email Connector module, that had seriously hampered overall product usability in case of unavoidable server-side crash.
    • Debugged and solved build issues with JNI exports in DLLs that caused intermittent application hang problem.
    • Thoroughly Analyzed the crash reports and provided sufficient evidence to prove the issues with 3rd party components/APIs.
    • Debugged ANT based build structure to resolve build issues.

WebMeeting (Screen Sharing Application) for MAC and Windows:

  • Role: Backend Developer (26 Months)
  • Accomplishments:
    • Received “You Made a Difference Award” for the extensive work done in the initial phase of the project, which helped the team to scale up and gain Client’s Confidence.
    • Understood the Chromium Open-Source Project relevant modules, build systems which can be reused to create a cross-platform (Windows & MAC) Screen Sharing application which used VP8 video codec and WebM packetizer. Demonstrated the key functionality of screen sharing in 3 months.
    • Evaluated and AWS S3, Dropbox, Google Drive for file sharing capabilities, but Client did not pursue another cloud platform and file share was implemented with Firebase & PubNub.
    • Designed and developed HTTP transport module for screen sharing data + Chats + File sharing modules with Firebase & PubNub cloud platforms.
    • Later, it was developed into a full-fledged product with multiple browsers supports + Chat, file share, recording capabilities.
    • PoC application was developed for Image & Text Detection in screen share data using an open-source library.

    • The core screen sharing product WebMeeting is still in production.

Porting WebRTC based GChat Application on Linux & Android Platforms:

  • Role: Backend Software Developer (20 Months)
  • Accomplishments:
    • The product was successfully showcased in CES2012-13 by the customers. Responsibilities:
    • Set Top Box, TI BeagleBone, Panda ARM board bring up with Linux/Android OS images.
    • Building & Porting WebRTC code for different Linux, Android Set top boxes.
    • Dealing with all kinds of compile and run-time errors on every platform.
    • Modifying WebRTC GYP build structure for integrating platform specific Codec and Camera Libraries.
    • Integrating H264 Decoder-Renderer APIs with WebRTC code.
    • Integrating/Testing Quanta and Maxim Camera modules in the applications.
    • Debugged H264 Decoder, YUV-RGB Render modules for Linux, Android platforms.
    • Thorough Unit Testing & Bug Fixing.

IL&FS Science Exploriments:

Search