Md Saiful Islam (SAYEF)
Introduction
Erlangen, Germany · (+49) ASK-4-IT! · hello@sayef.tech
I am a tech lover, an AI fanatic, experienced in developing tech solutions in the field of artificial intelligence, especially in computer vision and natural language processing. I have also spent a quite a lot of time in web development contributing to some enterprise applications.
Skills
Tools
- PyTorch
- Keras
- TensorFlow
- Kubernetes
- Docker
- Gensim
- OpenNLP
- CoreNLP
- Matplotlib
- Scikit-Learn
- Linux
- Git
- ROS
- OpenCV
- Oracle Database
- MySQL
- PostgreSQL
- NoSQL
- HTML
- NodeJS
- Flask
- Scrapy
- jQuery
- Redis
- Solr
Programming Languages
- Java - Proficient in building web applications, REST APIs using Spring MVC, Spring Boot and EJB.
- C/C++ - Experience in ACM ICPC contest programming, ROS and image processing.
- Python - Skilled and machine learning experience as primary language. Also proficient in building REST APIs using Flask/FastAPI.
Experience
NLP Engineer
Fraunhofer IIS, Erlangen – (January 2022 - Present)
Developing tools and methods using MLOps to bring NLU research into production for the NLP team of Audio and Media Technologies group.
Research Assistant
Fraunhofer IIS, Nuremberg – (September 2020 - December 2021)
Collaborated in different NLP and knowledge representation tasks to identify, analyze and evaluate market and technology for the Future Engineering team.
Software Engineer (Werkstudent)
IAV (Automobile Engineering) GmbH, Gifhorn – (March 2019 - February 2020)
Contributed in different computer vision tasks for the UAV team in making drone-based security/surveillance systems.
Senior Software Engineer (AI)
TigerIT Bangladesh Ltd, Dhaka – (January 2019 - November 2021)
R&D in the fields of machine learning, deep learning, computer vision, image processing and natural language processing.
Software Engineer
TigerIT Bangladesh Ltd, Dhaka – (January 2016 - January 2019)
Software development of enterprise application.
Education
Georg-August-Universität, Göttingen (Germany)
MSc, Applied Computer Science – (September 2018 - September 2021)
Data Science with Computational Neuroscience Subdomain.
University of Dhaka, Dhaka (Bangladesh)
BSc, Computer Science and Engineering – (January 2012 - January 2016)
Projects
This section contains awesome projects that I’ve developed:
ANPR (Bangladesh)
Automatic number plate recognition system. (All rights to TigerIT Ltd.)
An automatic license plate recognition system, based on several deep learning models for each task, i.e. automatic license plate detection, finding text area, and recognition of city name, vehicle class and number simultaneously.
Language & Tools: Python, PyTorch, Flask, Deep Learning
FRS
Robust face recognition system for office surveillance. (All rights to TigerIT Ltd.)
A surveillance application system for recognizing intruders in corporate office premises. The deep learning framework PyTorch has been used with transfer learning from open-source trained models.
Language & Tools: Python, ONNX, Flask, PyTorch, Deep Learning
EmoCam
Emotion Based Video Recommendation System. (Personal)
An approach towards video recommendation system based on emotion read from the face of a user in front of the camera.
Language & Tools: Python, NodeJS, Keras
Corella
An NLP based Dialogue Engine. (All rights to TigerIT Ltd.)
Corella is the heart of AI-driven application KinderMate which is targeted to kids’ entertainment and knowledge. It synthesizes kids’ search query using NLP techniques and automates decisions to pull contents from predefined knowledge pools. It has also hand-crafted conversation mode that helps kids learning through conversation games.
Language & Tools: Python, Java, OpenDial, OpenNLP, Gensim, Scikit-Learn
DocRefiner
A document image refiner. (All rights to TigerIT Ltd.)
An automated image resizing and refining tool that aligns the skewed images of the documents as much as possible without losing any vital information i.e. writing, figures, signatures, seal, etc. It can adjust brightness by increasing intensities of semi-washed data. It also reduces file size without losing readability of the document.
Language & Tools: Python, Computer Vision, Keras, Neural Networks
OCR for Number Plate Recognition of Nepal EVR
OCR based automated QA of number plates. (All rights to TigerIT Ltd.)
The ultimate goal of this domain-specific application is a quality assurance of the vehicle’s number plate after emboss printing. The application can detect zone name, plate number, plate color, font color, screw holes and SIT tag holes.
Language & Tools: C++, OpenCV
EVR Tajikistan & Nepal
Vehicle Registration System for Tajikistan and Nepal. (All rights to TigerIT Ltd.)
This enterprise solution allows people of Tajikistan (and Nepal with different requirements) to register vehicle information, capture owner’s photo, signature, fingerprints and upload other necessary documents. Later on, this information can be used to search the details of any vehicle and renewing the registration. This system can also find the criminal match for a vehicle.
Language & Tools: Java, Spring
3D Carrom Board Modeling
An OpenGL graphics project to modol carrom board.
A C++ project which was targeted to implement all basics of Computer Graphics i.e. modeling, lighting, shading and movements of objects with the help of OpenGL library.
Language & Tools: C/C++ OpenGL
99-Text-Compressor
A text file compressor and extractor. (Personal)
A C/C++ project to compress large text files and decompress them without any data loss using the most popular Huffman algorithm.
Language & Tools: C/C++
Awards / Certifications / Contests
- Computer Vision for Faces – By Big Vision LLC. Certificate earned at Wednesday, November 15, 2017.
- Neural Networks and Deep Learning – By deeplearning.ai on Coursera. Certificate earned at Saturday, December 23, 2017.
- Improving Deep Neural Networks, Hyperparameter tuning, Regularization and Optimization – By deeplearning.ai on Coursera. Certificate earned at Sunday, February 11, 2018.
- Structuring Machine Learning Projects – By deeplearning.ai on Coursera. Certificate earned at Sunday, February 25, 2018.
- Convolutional Neural Networks – By deeplearning.ai on Coursera. Certificate earned at Monday, March 26, 2018.
- Sequence Models – By deeplearning.ai on Coursera. Certificate earned at Sunday, April 15, 2018.
- Participant of ACM ICPC Asia Regional Dhaka Site (2013 & 2014).