**Kaggle achievements**:
- World’s first Quadruple Grandmaster
- Top 5 in all categories
- Kaggle profile: http://www.kaggle.com/abhishek

Get my book: Approaching (Almost) Any Machine Learning Problem for FREE: https://github.com/abhishekkrthakur/approachingalmost

I post a lot of cool stuff from time to time on my YouTube channel. Subscribe me on YouTube: https://www.youtube.com/AbhishekThakurAbhi

I hold a Master's Degree in Computer Science from the University of Bonn. In my free time, I like to take part in machine learning competitions and have taken part in over 200 competitions. My research interests include automatic machine learning, deep learning, hyperparameter optimization, computer vision, image analysis and retrieval and pattern recognition.

Abhishek Thakur

Interact Brands

Research on multi-modal models,  Creative x AI

AI Product

H2O.ai

Document AI - Kaggle Grandmasters Team

Developed multi-modal document understanding and layout language models 
Worked on inference optimization through tools such as, ONNX, Torchdynamo, nvFuser and TensorRT, leading ~2-5x performance (added a PR on Pytorch repository for dynamo)
Built a post-processing pipeline and template methods for better table extraction to boost the performance of the LayoutLM model
Built custom evaluation and integration testing framework for measuring performance

Data Scientist

MPAY CJSC

Manipulating, preprocessing, and extracting insights from large disconnected financial datasets
from PostgreSQL, SQL Server, and AWS
Implementing internal process improvements: automating manual processes with Python,
optimizing data delivery, re-designing infrastructure for greater scalability
Working with stakeholders including the Executive, and Product Leads to assist with data-related technical issues, supporting their data infrastructure needs, and presenting weekly reports

Data Engineer

Rəqəmsal İnkişaf və Nəqliyyat Nazirliyi - The Ministry of Digital Development and Transport

Using BI tools to extract value from data by means of data viz technics 
Analysis of data with the help of libraries of Python (Pandas, Numpy, Seaborn, Scikit-Learn)
Using NLTK and SoTA models to implement Natural Language Processing applications
Working on Computer Vision algorithms for Object Detection applications with models: YOLOV3, TF Object Detection API

Kaggle

Solving real-world Machine Learning problems from Google, Stanford, WHO, Lyft, etc.
Applying full scaled Data Analysis, Explatory Analysis, Feature Extraction technics with tools like Numpy, Pandas, Seaborn, Plotly, Sklearn, AutoML 
Developing accurate Machine Learning models  95+ (Top 2, 5, 10%) score with TFKeras and Pytorch

An engineer passionate about building algorithms that live up to theoretical guarantees on paper. Better, faster, and ultimately safe for humanity.

taplio

The Battle

Abhishek Thakur

Elvin Aghammadzada

Followers

Posts

Engagements

Likes

Ready to grow your LinkedIn presence?