Sign in

C,Linux kernel,python,embedded systems machine learning and deep lear

Secunderabad, Telangana, India
November 11, 2018

Contact this candidate


Professional Profile

** **** ***** ** ***** work from ground up gave me the opportunity to Data Scientist and embedded in the software industry involving AI, Machine Learning Algorithms,Statistical for a mid-sized and large scale IT services company.

Predictive Analytics, Regression techniques, Linear regression, Time series, Market Basket Analysis, Decision Tree Learning, Machine Learning techniques, Neural networks, SVR, SVM, Naïve Bayes, Random Forest, KRR, KNN,NLP and K-Mean Cluster.

Business forecasting and decision support for Statistical and mathematical analysis.

Identify new process or areas for improvement opportunities. Analyze, identify and interpret trends or patterns in complex datasets. Involve in data science, visualization, storyboard generation, data integration, and Descriptive analytics for AYS and CMIS data.

Search engine analytics: prediction of keyword conversion for keywords using Ngram model pull out from KB Article Knowledge based system.

CCM Customer Conversation Model Voice Crawling and text mining techniques to score call center Agent, generate keyword taxonomies, and assess commercial value of bid keywords Finding Competitor promotion

Design and development of Software, including Design, Enhancement and Maintenance of Test framework for embedded applications and products

Compiler design for GPU processors including linker, assembler and binary code generation

Programming and optimizing firmware from-the-scratch for SOC, sensors and Linux Kernel

Ability to provide leadership, motivation and strategic direction to the team with excellent interpersonal, collaboration, and problem solving skills

Experience working with both Agile and traditional development models

Hands-on technical experience

Hardware bring-up line AU1200, A2000, Beaglebone

Expertise in trailoring down system to the needed features

Expertises in automating build systems and running apps on embedded products in python

Expertise in understanding, fixing issues and optimizing build systems.

Rendered Onsite support for 3 months at Taipai, Taiwan for UAT and Final release at customer site

Technical Skills

Languages: C

Operating Systems: Linux, Contiki, VxWorks

Debuggers: Trace32, openocd

Analytics & ML: Python, Tableau

Supervised: Linear Regression, Logistic Regression, Neural Network, SVM, Decision Trees

Un Supervised: K-means clustering, Anomaly Detection, Recommender Engine

Databases: MySQL

Protocols/Libraries: PTP, Numpy, Pandas, scikit-learn

Version Control: Gerrit, CVS,SVN

Work Experience

Senior Member Technical Staff, Ineda systems Pvt Ltd, March 2014 - June 2017

Integrated Open Source modules for developing firmware for external sensors

Evaluated existing system for porting to ThreadX OS

Defined and implemented key performance indicators (KPIs) for the wearable devices

Defined process for creation and validation of SDKs from an end user perspective

Lead Technical Consultant, Accel Frontline Limited, Nov 2011-March 2014

Handled technical and management part of the project

Recruited resources and trained them to work in various modules of the compiler.

Reported project status, handled online weekly status meetings with customers

Managed team to resolve day-to-day issues with emphasis on SDLC

Delivered phase wise product delivery in committed scheduled date

Bring up of Beaglebone as per customer requirements and implemented OCR module of opencv.

Senior Technical Associate, Tech Mahindra Ltd Pune, September 2005 – March 2009

Leading the project

Requirement elicitation, enhancement, maintenance review and delivery

Software Engg, CNetworks Consulting Services, Bangalore,

July 2000 – July2004

Creating Test specification

Blackbox testing

Porting product from across OS

Creating Unit test based on the specification

Academic Details

B.Tech in Electronics & Telecommunications Engineering, National Institute of Technology, Silchar, formerly REC (Regional Engineering College), Silchar, 2002

Completed Japanese Language Basic Conversation Course from ABT AOTS DOSAKAI

Project-Wise Experience


Dynamic Advertisment Insertion for live broadcasting


July 2018 – Till Date


ML Engineer


Develop technique for creating hash tables in the form of signature for frames procured from MPEG-TS stream. To compare signature of incoming advertisement with the signature of the advertisement present in the database in GPU and advice user of replacing remote advertisement with regional advertisement.


To split frames from MPEG-TS into images for generating keys using MD5 and SHA3

To transfer configurations and keys from server to client

Build the ML model using deep learning, reinforcement learning, face detection using CUDA and keras libraries.

Owned the design and development of analyses, dashboards, etc., to drive key business values.


Conversation Analyzer


July 2017 – July 2018


Machine Learning Engineer


Developed two machine learning algorithms (using Natural Language Processing) to automate the process of evaluating the questions asked by customers to a chat bot.


As a Machine Learning Engineer Worked on Natural Language Processing (NLP) . Performed tasks such as feature extraction Stemming, Synonym replacement using nltk library of python (mostly TF-IDF)

Automated the process of marking of conversations of a chat bot using two machine learning algorithms (ML)

Build the ML model for mapping of customer questions to the key questions (intents) from scratch. Carried out all steps like - data cleaning and it's preprocessing.

Explored the whole data using R in R Studio and pandas library in python.

Used the sklearn library of python for using various already defined machine learning algorithms


Data Analysis and Interpolation


Sept 2015 – March 2017




Data Received from wearable device via BLE served as input for analysis using machine learning for Pedo, HRM and weather. For record purpose one week data is taken into considerations and model is built around data after every week in a continuous automated process.


Developed strategy and road maps including tactics, project plans, milestones and KPIs.

Carried out all steps like - data cleaning and it's preprocessing.

Created visualization dashboards for weather and PEDO data.

Merging needed variable using PCA leading into selection of features.

Building ML model needed using various regression algorithms.

Automating process of getting ML model and predictions.


Development of Wearable device


March 2014 – June 2017


Senior Developer


This wearable device has various features like RTC, Pedo, HRM, Weather, BLE, Settings screens. It runs both on MCU and MPU. Sensors run on MCU to control power consumptions and media player WiFi. Others run on MPU. It has raw bootloader which can boot both MPU and MCU. It is supported with TCM, SRAM and flash memories.


Worked for integrating open source modules into existing modules with developing firmware needed for external sensors

Porting of CFS filesystem for wearable devices for storing metrics related to Pedometer, HRM and BLE along with optimizing linker scripts for optimizing usage of SRAM and flash memories

Evaluating existing system for porting to ThreadX OS

Conversion of script based build system to gnu makefiles based build system along with optimizing build system to support debuggers like Lauterbach and Bus blaster

Defining and implementation of needed key performance indicator for the wearable devices in Python.

Fixing Issues with the stability of the system along with enhancing product for the dynamic loading of the module at run time

Defining process for the creation of the SDKs and validating SDKs from end user perspective in python

Development and automation of the tools needed for the validation of the SDK for the wearable devices in python

Implementing Gerrit along with GIT for the version control


Shader compiler for GPU


December 2011 – March 2014


Project Lead


This is Shader compiler for Embedded GPU processor. It converts the input shader language files (vertex/fragment) in to the Target GPU architecture assembly and binary code.

It has two phases: Front End Processor and Back End processor. Front End processor converts the input shader file in to intermediate form (internal format). The Back End processor reads the input intermediate form and transfers it into target assembly and binary codes.


Preparing HLD and LLD

Setting up complete development environment for the entire project life cycle: setting up users and tools, various environment variables, configuring CVS etc.

Handled the project from Design to Maintenance Phase across various modules with a team of 17 members


Lex parse

Support of architecture specific functionalities using Pragma

Syntax, Semantic and Expression evaluation

Built in function development

oDesigning intermediate format (Binary) to pass the tree structure from FrontEnd to BackEnd through file


Packing (memory)

Code Generator

Linker and Binary Generator

oSupport of Online and Offline compilation

Release Manager for incremental and final release

Bring up of beaglebone with tailored down Linux OS

Implementing printer apps to print images using avahi feature on beaglebone in python

Development of OCR application for the extraction of text from images and videos using opencv.

Applying neural network and bagging concept on images to verify using machine learning approach for OCR in python


OS Abstractor Product validation on different Target RTOS’s


July 2010 – July 2011


Project Lead


OS Abstractor is a C/C++ resource virtualization framework that provides a flexible and robust real-time application development which prevents software from being locked to a specific operating system and a specific version. Can easily port and run on host machine and optimize legacy code to multiple target platforms.


Developing mechanism to avoid system deadlock using Resource Control Block concept.

Worked on developing test applications to validate the functionality

Preparing plan for virtualization at the resource level to optimize usage of resources.

Validated Thread, process, handle, timer, mutex, semaphore, and event related APIs by developing test application by using existing test frame work.


Maintenance and Enhancement of Vision Map Maker


April 2009 –March 2010




Task of maintaining product and need to do require upgrading that would keep our product in the market with latest requirements which is measured on FPGA so as to get profiling and performance measure of the product in the real time.


Procurement for various designs and profiles to analyze the product on FPGA and checking the power consumption of each module in detail

Conducting brain storming sessions and lesson learnt meetings with other teams

Customer Communications as SPOC


Portable media player accessory


October 2005 – March 2009




Linux 2.6.11




The development of the stack, which is primarily used for printing images on Pictbridge compatible printers. Camera and Image Processing Association (CIPA) standard which is for printing images with are of minimum resolution 640 x 480. The stack developed can be loaded on to the new kernel so as stack placed on AU1200 will interact with printers. Media player has been extended for the audio and video file playing at the run time.


Project management and Release planning for product software.

Identify technical solutions, workarounds and resolve resource allocation

Developing design high level and low level software design

Developing different drivers required for the implementation of the product

oConversion of mjpeg to jpeg.

oEncryption and Decryption.

Interacting with the client from the design phase through final release.


Developing Application for printing image transformation


August 2004 – September 2005




Camelot, Photoshop

Role/ Designation



The printer driver that prints image using RGB format do print images in black and white. Here we are having server application, which waits for client application for the input of the file and configuration parameters of the image. Once application gets them it changes the image into raw format using post-scripting application that is running the background. Now the raw format is again into TIFF format for getting image tuned into CMYK format, which is printed.


Developing detailed WBS required for accomplishing task

Developing design high level and low level software design.

Developing TIFF writer required for the application

Interacting with the Client from Requirement Gathering phase to the Final Release


Porting and adaptation of Single Tasking Softplc.


June 2002 – July 2004


ISagraf, CodeSys, VxWorks




It is designed to be system-independent, thus enabling rapid adaptation of the CoDeSys programming system to any given PLC. It implements the complete functionality of a PLC cycle and supports communication with the programming system as well as all debugging functions that are supported by the CoDeSys programming system.


Black Box Testing

Creating Test specification

Creating Unit test based on the specification

Creation of Test suite for testing product features


Porting of Internet Camera application


August 2000 – May 2002


eCos, GCC

Role/ Destination

Team Member


Implementation of the IP camera using MPEG-4 standards. This involves porting to eCos operating system from Windows-NT the application and then implementing the GA-4004 protocol stack on it for online viewing picture live at required instances.


Developing the GA-4004 from MPEG-4

Porting the application from Windows to eCos

Testing the credibility of the Application

Commit the solution in the git once the given solution is reviewed.

Personal Details

DOB: 25th August 1977

Passport Number: H7049232

Address: H.No.32-77-32/3,DN-12

Devinagar 1st Street,

Neredmet, Hyderabad - 500056.

Contact this candidate