Hi There!
I'm MANISH SONI
Camera Software Application Engineer at Qualcomm.
Graduated from Michigan Tech.
An independent thinker. Intelligent Autonomous systems enthusiast.
- Independent Thinker
- MS in Electrical Engineering
Interest Area: Machine Learning, Computer Vision and IoT.
My interest area revolves around sensing and processing for intelligent autonomous systems. I use Machine learning, Computer Vision, IoT, Firmware and Software Engineering tools to implement and craft the state of the art solutions.
My Expertise
Software Engineering
3+ years of professional experience as a System Software Engineer
I am a Polyglot Programmer. I have used various language to complete the projects.
Polyglot stack: Java, Python, C, C++
IoT & Firmware
Have worked with UEFI Firmware and Project related to thin clients.
Debugging
A developer who developed a debugger.
Expert skill level.
Profiling, debugging, Inspection and logging say a lot.
Web Services / REST-API
Used Eclipse Jersey, Spring boot, Flask framework to implement the consumable rest apis.
Machine Learning
I got fond of Machine Learning after coming to USA. My academic courses and projects pushed me closer to this field.
Relevant Courseworks: Machine Learning, Computational Intelligence, Artificial Intelligence.
Computer Vision
Have implemented a number of projects using image processing and computer vision.
Opencv, MATLAB and CNNs.
What I Did
PROJECTS
01.
Minimal Human intervention Billing System using Keras Retinanet and TensorFlow
Used state-of the art machine learning methods and tools to implement the CNN solution which could generate the bill using one single snapshot of the products laid at the countertop tray.
02.
Ideation & Implementation of Cost-Based GPSR protocol for improving the PDR, Latency and Throughput in VANETs for V2V communication.
Actively researched and ideated a better approach for improving the perimeter mode of GPSR based on cost parameters which improves replaces default approach of using a right-hand rule.
03.
Leader Election Protocol for fault tolerance and adaptation in Low power IoT system.
Implemented an effective leader election protocol along-with neighbor discovery protocol to discover the low power distributed IoT devices in Mesh/Line/Ring/Tree topology that can accommodate a new device at real-time basis and could elect a leader having fault tolerance and adaptation.
Most Recent
Project
Ideation & Implementation of Cost-Based GPSR protocol for improving the PDR, Latency and Throughput in VANETs for V2V communication.
