profile picture

Satyasheel

Data Engineer | MLOps | GCP

Education

MSc in Machine Learning

Royal Holloway University of London
2015 - 2017

BE in Electronics & Communications

Birla Institute of Technology, Mesra
2008 - 2012

Honors & Awards

Top 20% , Bosch Production Line Performance (298/1373)

Kaggle
2016

Big Data Scholarship

Royal Holloway University 0f London
2015

Finalist, Nasa Lunabotics Mining Competition - Team Grutuva

NASA, Florida
2012

Talks

2018

Apache Airflow in the Cloud: Programmatically orchestrating workloads with Python

PyData Conference London

2016

Get your hands on implementing a Flink app

Flink London meetup

Certificate

Google Certified Profession Data Engineer

Udacity Self-Driving Car Nanodegree

Certificate in Engineering ExcellenceinBig Data Analytics and Optimization

Technical Skills

Languages

Python, Scala

Technologies/os

Apache Spark, Apache Airflow, Apache Beam, BigQuery, Google ML Engine, PyTorch, Keras, Git, Linux

Cloud Platform

Google Cloud, Azure

Career Profile

I am a Machine Learning Engineer with experience of developing end-to-end machine learning/data pipeline with data validation and pipeline testing. I have lead the design and development of Data Ecosystem on GCP as well as Azure for different client.I am a certified Google Cloud Data Engineer and also a nanodgree holder of Udacity Self-Driving car. In my free time I always look out for learning something new or build something new.

Experiences

Data Engineer Manager

Nov, 2019 - Present
Quantexa Ltd, London

  • Building scalable data pipeline on Google cloud for anti money laundering system using Scala, Apache Spark, Elastic-search and Apache Airflow.

    • Machine Learning Engineer

    July, 2016 - Nov 2019
    Data Reply, London

  • Lead the design and development of MLops pipline using Sagemaker, MLFlow and S3 for one the leading sports retailer in the world
  • Lead the design and development of Data Ecosystem for on of the leading retailer in the UK on Azure
  • Developed a Machine Learning pipeline for an online retail organisation in the UK, predicts which customer is going to churn in next 90 days. This whole pipeline was built on Google Cloud and orchestrated via Apache Airflow
  • Built a predictive and analytical pipeline for state agency in Ireland which identifies wether an operator is skipping or manipulating the vehicle test with tester.
  • Built a demo called "celebrity face match" using PyTorch (Face Embedding - Google) and Flask (for calling the api for inference)
  • Developed multiplayer pong game over the network which identifies the human hand gesture (Using Convolutional Neural Network) and plays the game. Model which Identifies the hand gesture was trained on custom generated Images.
  • Developed an analytical pipeline which performs number of ETL operationas and the update a visualisation dashboard for an analyst to perform further inspection of delevery and purchase data. This pipeline was deleoped for an Online Retail organisation in the UK.

    • Data Scinetist

    December 2014 - August 2015
    Cyient-Insights, Hyderabad, India

  • Built an analytical pipeline for Identifying the faulty sensor or going to be faulty sensors of an Aircraft for a Civil aviation sector organisation, tools I used were R and a single system with 32 GB RAM plus 16 Core Processors.
  • Mentored two interns, this includes helping them to get the feel of real world data science and how to productionize a data science project.
  • Created a training workshop for fresh graduates in the organisation, this training inclides Machine Learning with python.
  • Trained 11 graduates on Data Science/Machine Learning with Python

    • System Engineer

    Ocotber 2012 - November 2014
    Business Process Management Team, Infosys, Hyderabad, India

  • Developed an business process for large scale automated document verification. This process was developed for a leading financial organisation in USA.
  • Gave training on Appain Business Process Managment to new graduates joing Infosys BPM team.

    • Markov-Chain for document ranking - Used markov chain simulation to rank list of emails in a tree network. This project was written in Java.
    Neural-Network form scratch - Native python based two layer neural-network from scratch. All optimisation and gradients calculations written from scratch
    K-NN Classification from scratch - Developed and implemented multi-class image classification using euclidian and tagent distance metric.
    Finding Lane Lines - Efficient implementation of computer vision algorith to find and lable the lane lines ion road using open CV.
    Multiple Self driving car projects which I have done during my Udacity Self-Driving car nanodegree - Implemented multiple self-driving car project. Some of them are Sefl-Driving car PID control, Nonlinear Model Predictive Control with actuator delays, Localisation in self-Driving car
    -->

    Extracurricular Activites

    Organiser

    Jan 2016 - Present
    Flink London meetup

  • Organised several flink meetup with different speaker on topics like, streaming analytics, streaming sql and several Flink use cases.

    • Member

    September 2015 - April 2016
    Student-Staff committe at Royal Holloway University of London

  • Organised several meeting with staff in Royal holloway university of London to discuss the progress of different student and thier concern.