Banu Atav
Verified Expert in Engineering
AI Specialist and Developer
Rotterdam, Netherlands
Toptal member since August 24, 2020
Backed by a master’s degree in econometrics, Banu is an AI specialist with three years of industry experience in building data-driven automation solutions using machine learning and natural language processing. Banu has delivered over ten projects ranging from proof of concept and ideation phases to solution deployment. Skilled in various techniques such as entity extraction, text classification, sentiment analysis, text summarization, among others, Banu can deliver on your AI requirements.
Portfolio
Experience
- Generative Pre-trained Transformers (GPT) - 3 years
- Machine Learning - 3 years
- Python - 3 years
- Natural Language Processing (NLP) - 3 years
- Git - 3 years
- Statistics - 2 years
- Google Cloud Platform (GCP) - 2 years
- PyTorch - 1 year
Availability
Preferred Environment
Visual Studio Code (VS Code), MacOS, Git
The most amazing...
...project I've done is create a process that integrates both machine learning and human employees to improve both speed and quality of the completed tasks.
Work Experience
Freelance Data Scientist (NLP, ML)
Bayesia
- Designed the flow for detecting sections in contracts and matching sections against templates using information-retrieval / document ranking techniques (TFIDF and shallow neural networks: GloVe).
- Deployed a Python application as a Flask app on AWS (using Elastic Beanstalk).
- Augmented Microsoft Word to an HTML5 parser (Pandoc) with Python filters to improve block-line sectioning and list structures.
Freelance Data Scientist (NLP, ML)
Bayesia
- Composed for a project (that dealt with media bias detection in news articles) a relevant dataset and designed and set up a labeling process for labeled dataset generation that ensures labeling consistency and quality (using statistics like Kappa).
- Project:. Created for a project (that aimed at detecting media bias in news articles) a BERT model for detection of six bias types using text classification (HuggingFace, PyTorch).
- Created for a project (that involved a recommender system MVP on online recruitment) the MVP model architecture to provide the business with the right information on their data collection process in the future.
- Built a website summarization tool using T5 and BERT that scrapes website content using Beautiful Soup; also explored both abstractive and generative summarization techniques.
AI Lead | Machine Learning Specialist
Ciphix
- Developed a machine learning solution for the automatic processing of incoming IT service tickets.
- Created an entity extraction application that extracts information from text files for further automatic processing.
- Created an Azure machine learning pipeline for automatic retraining/updating for deployed machine learning models.
Data Quality Consultant
Dun & Bradstreet, Rotterdam
- Assembled a structured dataset from unstructured contractual information (it was text-based and involved multiple languages).
- Took the initiative to adapt this process into a scalable and automated solution by implementing OCR and PDF scraping algorithms.
- Transcribed 100s of the contractual data in Salesforce.
Experience
Web Page Summarizer
Scrape a web page from a provided URL then select the main text and deliver a summary.
Method:
The source code of the web page was parsed using Beautiful Soup with us exploring several summarizers. We had four different extractive summarisers using the following packages Gensim, Summa, NLTK (Natural Language Toolkit), and spaCy. We then implemented two pre-trained abstractive summarisers using HuggingFace's transformers package (a wrapper around PyTorch and TensorFlow).
Legal Contract Matching
Media Bias Detection in News Articles
Education
Master of Science Degree in Econometrics
Erasmus University Rotterdam - Rotterdam, Netherlands
Certifications
Deep Learning Specialization by Deeplearning.AI
Coursera
Full-stack Web Development with Flask
Pirple
Machine Learning with TensorFlow on Google Cloud Platform
Coursera
Skills
Libraries/APIs
TensorFlow, PyTorch, NumPy, Pandas, Scikit-learn, REST APIs
Tools
Git, MATLAB, Azure Machine Learning, Gensim, Amazon Elastic Block Store (EBS), Boto 3, Pandoc
Languages
Python, R, HTML5, CSS
Frameworks
Flask
Platforms
Google Cloud Platform (GCP), MacOS, Visual Studio Code (VS Code), Amazon Web Services (AWS), AWS Lambda
Storage
Amazon S3 (AWS S3)
Other
Machine Learning, Bayesian Statistics, Statistics, Time Series, R Programming, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), UiPath, Information Retrieval, GloVe, Deep Learning, Transformers, Word2Vec, Computer Vision Algorithms, Neural Networks, Recurrent Neural Networks (RNNs), Convolutional Neural Networks (CNNs), BERT, Hugging Face, Tf-idf
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
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring