Tushar Karayil, Developer in Frankfurt, Hesse, Germany
Tushar is available for hire
Hire Tushar

Tushar Karayil

Verified Expert  in Engineering

Deep Learning Developer

Frankfurt, Hesse, Germany
Toptal Member Since
June 8, 2015

Tushar is a skilled software and machine learning developer with deep learning, data mining, and system programming expertise. He holds a PhD in machine learning and has seven years of software development experience in Python and C++, working for IBM Software Labs and Novell, Inc. He is an expert in deep learning systems with TensorFlow and Keras. He is a great communicator and a talented programmer.



Preferred Environment

Git, Python, Linux, AWS Lambda, Amazon EC2, Amazon DynamoDB, AWS IAM, Google AI Platform

The most amazing...

...thing I have researched is changing Transformer architectures to include natural language style information.

Work Experience

Machine Learning Developer

2012 - 2014
University of Kaiserslautern, Germany
  • Developed LSTM-based recurrent neural network frameworks.
  • Built image-generating software for synthetic image generation.
  • Programmed software for object detection and tracking.
  • Contributed to data mining projects with IPython and Pandas.
  • Created unit testing frameworks in Python.
Technologies: Decision Trees, Neural Networks, Statistics, IPython, SQL, MongoDB, Pandas, NumPy, Matplotlib, Python, C++, C

Senior Software Engineer

2010 - 2012
Novell, Inc.
  • Supported directory service software on SuSE Linux.
  • Created a testing framework for testing eDirectory on SuSE Linux.
  • Supported the distributed computing environment on SuSE Linux.
  • Developed an integration module for connecting directory services with other services.
  • Created the build framework for generating weekly builds.
Technologies: Python, Linux, C++, C

Software Engineer

2005 - 2010
IBM India Software Labs
  • Developed a communication framework for a middleware transaction manager.
  • Developed monitoring modules for the transaction manager for real-time monitoring.
  • Created testing frameworks for regression tests in Python.
  • Created stress test frameworks in Python.
  • Supported the whole product during releases.
Technologies: Python, Linux, C++, C

Optical Character Recognition for Cursive Unsegmented Character

OCR (Optical Character Recognition) for cursive characters is a challenging task for any machine learning researcher. Recurrent neural networks have demonstrated promising results on cursive connected characters. This project aims at using neural recurrent networks for unsegmented character recognition from images.

Object Detection and Tracking

Developed a module to detect specific objects inside images and track them across long video sequences. The module is used for tracking markers in crash tests of automobiles.

Movie Genre Analysis (Machine Learning/Data Mining)

Developed a python framework to extract main keywords from movie scripts. This data is then used to represent the movies as a Bi-Partite graph. These Bi-Partite graphs are later used for cluster analysis to see if similar genre movies form a cluster

Text Detection in Natural Scenes Using MSERs (Machine Learning/System Programming)

Python/OpenCV project to localize text in natural images and videos using Maximally Stable Extremal Regions (MSER). A demonstration of this project can be found at [https://www.youtube.com/watch?v=PiYfxU68k3U]. (tried through a 8 mp mobile phone camera)
2016 - 2020

Ph.D. Degree in Artificial Intelligence

University of Kaiserslautern - Kaiserslautern, Germany

2012 - 2014

Master's Degree in Machine Learning, Computer Science

University of Kaierslautern - Kaiserslautern, Germany

2001 - 2005

Bachelor's Degree in Computer Science and Engineering

National Institute of Technology - Calicut, India


TensorFlow, Keras, SciPy, Matplotlib, PySide, NumPy, PyTorch, Pandas


IPython, PyDev, GCC, Git, Subversion (SVN), AWS IAM, Google AI Platform


C++, Python, C, SQL, Bash


Linux, AWS Lambda, Eclipse, Amazon EC2


Data Science, Agile Software Development


Google Cloud, NoSQL, MongoDB, Amazon DynamoDB




Google BigQuery, Deep Learning, Machine Learning, Artificial Intelligence (AI), OpenAI GPT-3 API, Chatbots, Statistics, Neural Networks, Decision Trees, Software

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.


Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring