menu

Shyam Sundar Ravikumar



Machine Learning | Deep Learning | Computer Vision




About


An Associate Software Engineer at Collins Aerospace, Raytheon Technologies working on the Airbus Aircraft programs, and an Electrical Engineering Graduate from USC interested in Machine Learning, Deep Learning, and Computer Vision frameworks and keen on working on developing similar technologies.

Machine learning systems are revolutionizing the way we live, work, and interact with Deep Learning and Computer Vision being some of the key domains that are driving this transformation. I intend to learn and grow with the advancing technology trends.

A little about myself, I hail from Coimbatore, India, and I am a Shutterbug, a Formula 1 fanatic, a Travel buff and a Gastronomist (Would love to have Dosa anytime!).

profile_pic


Experiences

company_vector

Collins Aerospace, Iowa

Jun 2020 - Present

Associate Software Engineer

Working on developing software for the Airbus Aircraft programs.


python logo
robotframework logo

Frameworks: SVN, Agile Methodologies

chevron_rightLearn More
company_vector

Full-Stack Developer Intern / Co-op

Developed a Flight Data Simulator that emulates sensor data from the Aircraft Flight Control System and integrated it to the Software Stack to simulate the Moving Maps Software Airshow-2 in Real-time.

Worked on the Software development cycle of the Next Generation of Interactive Moving Maps Software Package Airshow-3.

Reduced Software Testing time by 25% through Static Analysis to identify vulnerabilities and enforce better coding practices.


Frameworks: OpenGL, SVN, Agile Methodologies

chevron_rightLearn More
university_vector

Indian Institute of Science, Bangalore

Jun 2018 - Dec 2018

Project Assistant

Member of the Sub 6GHz Baseband group working on the Indigenous 5G Testbed Project of India under the guidance of Prof. Chandra R Murthy, Prof. Neelesh B Mehta and Prof. K V S Hari.

Responsible for contributing towards the development of PHY and MAC layers, primarily focussing on the transport channel blocks and benchmarking performance for further integration to higher layers of the protocol stack.

I also developed the official webpage for IISc's 5G-Testbed.


Software Stack: OperAirInterface LTE and 5G-NR Software Stack

chevron_rightLearn More
university_vector

Indian Institute of Technology, New Delhi

Dec 2017 - May 2018

Research Assistant

Developed an Automated system for Translation of Prefabricated Nanostructured Sensors over 3 dimensions under the guidance of Dr. Anuj Dhawan, as a replacement to the microscope stage of inVia Qontor Confocal Raman Microscope. The system is designed using LabView and a GUI is developed to enable the user to control the translation of the stage.

Integrated a Temperature Control System with the Automated Translational Stage for Seamless analysis of Nanostructured Sensor Substrates. LabView is used to design the system and the GUI gives the user various modes to control the temperature and also allows the user to specify the wait times for the spectroscopic measurements at different patches of the sensor substrate.

Hardware: inVia™ Qontor® confocal Raman microscope

university_vector

Indian Institute of Technology, Chennai

May 2017 - July 2017

Research Assistant

Developed a complete OFDM system using the concepts of OFDM, Multicarrier Communication, Different FDM Techniques and Channel Estimation using Interpolation under the guidance of Dr. K. Giridhar

Successfully tested a modification in the LTE Frame Structure for Ranging in the PRACH aimed at efficient utilization of the available Resource elements for PRACH and allocation of unused Resource Elements for the User data.


Frameworks: Wi-Fi (802.11ac), WiMax (802.16)


Skills


Programming
  • Python star star star star star
  • C++ star star star star star
  • Java star star star star star
  • SQL star star star star star
  • Embedded C star star star star star
  • LaTeX star star star star star
Frameworks
  • PyTorch star star star star star
  • OpenCV star star star star star
  • Robot Framework star star star star star
  • NumPy star star star star star
  • Pandas star star star star star
  • Scikit-Learn star star star star star
  • Matplotlib star star star star star
  • TensorFlow star star star star star
  • ConvNets star star star star star
  • RNN / LSTM star star star star star
  • SciPy star star star star star
  • Seaborn star star star star star
  • Plotly star star star star star
  • OpenGL star star star star star
Technologies
  • Android Studio star star star star star
  • Git star star star star star
  • JIRA star star star star star
  • HTML star star star star star
  • CSS star star star star star
  • MATLAB star star star star star
  • LabVIEW star star star star star
  • Photoshop star star star star star
  • Lightroom star star star star star
  • Premiere Pro star star star star star
  • Arduino star star star star star
  • Agile Methodologies star star star star star
  • Bootstrap star star star star star
  • Unity3D star star star star star
  • Docker star star star star star

Projects


Amazon Whole Foods Delivery Slot Checker

  • Developed an automated script that identifies available Delivery slots from Amazon Whole Foods at times of high demand for groceries during the COVID-19 pandemic.
  • Supports Google Chrome and Mozilla Firefox Web browsers and notifies about the delivery slots (if available) through e-mail.



  • chevron_rightView Project


    delivery_slot


    Song Genre and Hit Prediction

  • Constructed custom datasets by extracting songs and its features
  • Trained different Machine Learning models for the Classification problem of Categorizing Genres and predicting the possibility of the song becoming a Hit
  • Analysed the performance of models on Validation data to select the best performing model
  • Evaluated and displayed results of the best chosen models on unseen data (songs)

  • [ML Algorithms]


    chevron_rightView Project




    genre_and_hit


    2D to 3D Reconstruction of Furniture Objects

  • Furniture objects segmented, cropped and centered through YOLO v3
  • Constructed a Model-free method using Autoencoders for 2.5D Prior Estimation and 3D Generation of models in Voxel Space
  • Additionally implemented a Model-based approach to Reconstruct 3D models with a reduced set of 15 parameters
  • Provided a comparative analysis between the performance of the two models
  • An Augmented Reality application was created to visualize the generated 3D models in a realistic environment



  • chevron_rightView Project



    2d3d_recon


    IDEAS - Intelligent Driver Enhanced and Augmented System

  • Winner of Ford Prize for the best use of Ford SDL API at CalHacks hosted by UC, Berkeley
  • Estimated Driver Emotion and Drowsiness (GCP Vision API) and closest Rest spots (Google Maps API) and queried to/by (JSON Requests) the Ford SDL API
  • Voice and Text Alerts are given if the driver is drowsy and 3 closest Rest-Spots are displayed on the Infotainment System
  • Based on Driver Emotion, a playlist of songs of similar mood is generated (Spotify API) and top 10 results are suggested on the Infotainment System

  • [Ford SDL]


    chevron_rightView Project





    Ford_Infotainment


    Human Activity Recognition

  • Worked on UC Irvine's AReM dataset to predict the Activity of a person based on Multisensor data
  • Analyzed the performance of Regression and Naive Bayes' Classification methods on the dataset and obtained a loss of 0.105 using L1 penalized Multinomial Regression

  • [ML Algorithms]





    Human_Activity_Recog


    Stinger - USC Self Driving

  • Lane Detection and Processing for a Self Driving Fully Autonomous Car Prototype
  • Employed Image Processing and PID control for smooth steering action and Speed Control
  • Demonstrated the functioning on a 1/10 scale RC car controlled using a Raspberry Pi 3

  • [Control Systems]


    chevron_rightWatch Video


    Stinger


    Multi-terrain Unmanned Vehicle

  • Implemented effective PID control mechanisms for stable Flight motion and Drift Correction
  • Signalling done through two channels to control the rotors and two channels for ground motion

  • [Flight Control System]


    uav



    Soil Moisture Estimation using Reflected GPS Signals

  • Proposed a Non Invasive solution for estimating soil moisture using a Software Defined Radio configured through MATLAB
  • Demonstrated comparable results derived through Invasive Soil moisture estimation methods


  • soil



    Virtual DrumKit

  • A virtual reality based learning environment built using Unity 3D, Arduino and motion sensors to help users learn to play percussion instruments
  • Data captured and processed by Android application modeled and programmed using Unity 3D Game Engine



  • chevron_rightWatch Video


    vr





    Coursework

    University of Southern California
    EE 503 Probability for Electrical and Computer Engineers
    EE 510 Linear Algebra for Engineering
    EE 559 Mathematical Pattern Recognition
    EE 660 Machine Learning from Signals: Foundations and Methods
    EE 569 Digital Image Processing
    CSCI 566 Deep Learning and Its Applications
    DSCI 553 Foundations and Applications of Data Mining

    Coursera
    Andrew N G, Stanford University (LA3MJVYAY7HZ)


    keyboard_arrow_up