Ronak Khunt, Developer in Amsterdam, Netherlands
Ronak is available for hire
Hire Ronak

Ronak Khunt

Verified Expert  in Engineering

Machine Learning Developer

Location
Amsterdam, Netherlands
Toptal Member Since
April 2, 2020

Ronak is an experienced data scientist and software engineer. His background includes building machine learning models, deploying them in productions with bare-metal machines or Docker. At Booking.com, he built a review moderation classifier model that helped classify millions of reviews and predictive models to strengthen their position in a competitive market. His expertise includes building robust data and prediction pipelines.

Artificial Intelligence (AI)Machine LearningGitPythonJavaScikit-learnKerasPandasNatural Language Processing (NLP)Deep LearningDockerNumPyMySQLJavaScriptNode.jsH2O

Portfolio

Toptal
Python, Scikit-learn, TensorFlow, Keras, Pandas, NumPy, Matplotlib, SpaCy...
Booking.com
Tableau, Spark, TensorFlow, H20, Scikit-learn, Keras, Python
Zilingo.com
Neo4j, Elasticsearch, Hue, Scala, TensorFlow, Python

Experience

Python - 6 yearsScikit-learn - 4 yearsMachine Learning - 4 yearsKeras - 3 yearsDeep Learning - 3 yearsData Science - 3 yearsTensorFlow - 3 yearsSpark - 2 years

Availability

Part-time

Preferred Environment

Git, Trello, Jira, Slack, MacOS, Linux

The most amazing...

...thing that I've built is a robust review moderation classifier that automated the moderation of millions of reviews each month, saving resources for many.

Work Experience

Machine Learning Engineer

2020 - PRESENT
Toptal
  • Developed a multi-class classification model to categorize promotional emails into eCommerce and transactions for a product company. Deployed the model to predict on batches using Docker.
  • Built a classification model to categorize if a domain is alive or dead based on the text of the domain for a product company. Deployed the model using Flask to be consumed by APIs. This model helped re-classify over one million domains.
  • Assisted in the identification of the right approach to the problem and identifying and cleaning up the dataset and feature engineering of the data. Built a complete pipeline for the data science problems.
Technologies: Python, Scikit-learn, TensorFlow, Keras, Pandas, NumPy, Matplotlib, SpaCy, Natural Language Toolkit (NLTK), Docker, Kubernetes, Jupyter Notebook, MySQL, Machine Learning, Deep Learning, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), GPT, Data Science, Artificial Intelligence (AI)

Software Developer (Machine Learning)

2018 - PRESENT
Booking.com
  • Built a review moderation classifier model to classify reviews, ensuring they do not contain offensive language and abides by the policy of the company.
  • Developed a model to strengthen our position in the competitive pricing market.
  • Created a model to predict whether the user will cancel the reservation or not. This is very helpful when providing the user with benefits.
  • Built an data pipeline to visualize a conversion funnel for the user journey which spreads over multiple days using (Py)Spark and visualization using Tableau.
Technologies: Tableau, Spark, TensorFlow, H20, Scikit-learn, Keras, Python

Software Developer (Machine Learning)

2016 - 2018
Zilingo.com
  • Classified images by identifying the person or object in the image and checking the uniformity of the background around the detected object. When the background is more uniform, the image shows better when displayed on a website.
  • Built a product recommendation system (similar products and products bought together) based on the data of the order.
  • Built a core library for a search-autocomplete for multiple languages using Elasticsearch.
Technologies: Neo4j, Elasticsearch, Hue, Scala, TensorFlow, Python

Senior Software Developer

2015 - 2016
NoBroker
  • Developed a framework to build an administrative dashboard without writing front-end code, having functionalities like search, sort, and filter.
  • Created a platform to build a rental contract between the landlord and tenant by filling out the details of the user and generating a PDF with possible customization.
  • Built a Node server to scrap data from various online sources by parsing an HTML response and storing it unstructured in NoSQL data storage.
Technologies: JavaScript, MySQL, Elasticsearch, Java, Python

Your Own Trigger Word

https://github.com/ronakkhunt/Trigger-Word-detection
Trigger word detection is used by many apps and has uses from mobile apps to home devices like Siri, Ok Google, Alexa, and many more.
Like Alexa or Ok Google, you can train the model to set any trigger word you like. For example, Jarvis.

Getting Started with Spark [Hands-on]

https://github.com/ronakkhunt/getting-started-with-spark
This notebook will allow you to start with (Py)Spark immediately without worrying about installation.
It contains details about how you can start using any data source to analyse your data with (Py)Spark.

Image Classification to Identify Bad Quality Product Images

Developed a system to identify bad quality images uploaded to an eCommerce platform then exporting them to third-party companies for improvement. The idea was to identify an object or person in the image and then calculate the uniformity in the background colors (in each quadrant of the image). If the background uniformity was lower than the threshold than the system would classify the image as one of bad quality.
2010 - 2014

Bachelor's Degree in Computer Science

Gujarat Technological University - Gujarat, India

AUGUST 2019 - PRESENT

Statistics for Data Science and Business Analysis

Udemy

APRIL 2019 - PRESENT

TensorFlow for Deep Learning with Python

Coursera

DECEMBER 2018 - PRESENT

Deep Learning Specialization

Coursera

MAY 2018 - PRESENT

Neural Networks and Deep Learning

Coursera

FEBRUARY 2018 - PRESENT

Machine Learning

Coursera

Libraries/APIs

Keras, Pandas, Scikit-learn, PySpark, TensorFlow, NumPy, Node.js, SpaCy, Natural Language Toolkit (NLTK), Matplotlib

Tools

Git, Tableau, Tableau Desktop Pro, Slack, Jira, Trello, Hue

Languages

Python, Java, JavaScript, Scala

Paradigms

Data Science

Platforms

MacOS, Linux, Docker, Kubernetes, Jupyter Notebook, H20

Storage

MySQL, Elasticsearch, MongoDB, Neo4j

Frameworks

Spark, Flask

Other

Artificial Intelligence (AI), Machine Learning, Natural Language Processing (NLP), Deep Learning, GPT, Generative Pre-trained Transformers (GPT), Text Classification

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.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

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