Md Iftekher Hossain

📍 Jyväskylä, Finland

📞 +358-7230756

👋 About Me

I am Md Iftekher Hossain, a dedicated data scientist with nearly 2 years of experience at SSL Wireless.Currently pursuing my Master’s degree in Cognitive Computing and Collective Intelligence at the University of Jyväskylä, I specialize in computer vision, machine learning, and AI model deployment. My passion lies in advancing technology to enhance our understanding of the world. I train my robots to perceive the beauty of our world and enhance their intelligence.

💼 Professional Experience

Data Scientist, SSL Wireless

Feb 2023 ‑ Aug 2023

Mentored juniors and interns.
Time series forecasting on daily usage data of a service and automatic measures on the prediction.
Developing automated invoice data parsing.
Developed a novel technique that detects whether a shop image is taken from an angled position

Junior Data Scientist, SSL Wireless

Nov 2021 ‑ Feb 2023

Developed shop image validator and shop appearance detector from image.
Developed a novel technique for shop size estimation from image.
Implemented custom segmentation model, which extracts shop portion from the whole image.
Developed an algorithm to detect objects of variable sizes with using only small object dataset.
Implemented FMCG Product Identification system for real retail stores in Bangladesh using computer vision.
Built and deployed multiple APIs in production using FastAPI and Docker.
Implemented NID Verification system with the accuracy of 98.5%.
Implemented a better face detection model for MFR (Multi Face Recognizer).
Implemented a novel technique called “Mean Embeddings” that increased the accuracy of the model by about 8% and speeded up by 5 more times than the previous method.
Built a robust Facebook Post and Comment scraper that scrapped over 60,000 comments from Facebook pages and groups.
Implemented a better Bengali voice recognition model with a Word Error Rate (WER) of 32.08% so far.

Vice President, Robo Mechatronics Association (RMA)

Sep 2019 ‑ Mar 2021 Robo Mechatronics Association Bangladesh (RMA) is the first university campus‑based robotics organization in Bangladesh. While working here we organized various competitions and workshops. Bangladesh.

Web Developer (Part‑Time), Panacea Private Consulting Group

Jan 2019 ‑ Mar 2019 Worked on a research and development project on Enterprise Resource Planning (ERP) using Django Framework.

Research Assistant, Institute of Energy Technology (IET), CUET

Mar 2018 ‑ Apr 2018 A research project for fabricating a Smart Water Meter Monitoring System using Python, MQTT, Microcontrollers, Sensors, and Raspberry Pi under the Institute of Energy Technology (IET), CUET.

🎓 Education

M. Sc. in Cognitive Computing and Collective Intelligence, University of Jyväskylä

Aug 2023 -

B. Sc. in Mechatronics and Industrial Engineering, Chittagong University of Engineering and Technology

Feb 2016 - Mar 2021

💻 Projects

FastAPI CRUD with Basic Authentication Developed a FastAPI application that supports CRUD operations with integrated Basic Authentication. The system prioritizes secure data access through protected endpoints.
A Home Security System with Social Interaction. A complete home security system is designed to act as a digital security guard. Face recognition, Voice recognition, Voice synthesis, and IoT technologies are used to make the system robust and user‑friendly. Faces are recognized using the Facenet model and a technique applied here to make the system faster and more accurate, called “Mean Embeddings”. The webserver was designedusing the Django framework.
Smart Eye Controlled Wheelchair for Disabled Person. Designed a wheelchair controlled by eye movements, enabling disabled individuals to move around easily and independently usingsimply by their eye movements. Color Sorting with Computer Vision. Developed an advanced computer vision system capable of accurately identifying and sorting objects based on their color variations.
Face Recognition using Mean Embeddings. Developed a cutting‑edge facial recognition system leveraging Facenet. A novel method, ”Mean Embeddings,” was introduced, resulting in a significant performance boost—achieving a 6.5 times speed increase and enhancing accuracy by 1.6 Face Recognition‑based People Counter on Raspberry Pi 4 Developed a system that accurately counts individuals using the Facenet architecture for face recognition. This innovative people counter has been efficiently implemented on a Raspberry Pi 4.
Conveyor Object Detection using OpenCV Implemented a real‑time detection system that identifies objects on a moving conveyor based on their shape, size, and color. Leveraging the capabilities of OpenCV, the project efficiently classifies and analyzes objects, optimizing conveyor operations.
Intelligent Voice Assistant using Rasa Framework. Engineered a real‑time voice assistant that provides intelligent responses. Utilized the advanced capabilities of the Rasa framework to ensure dynamic and interactive user experiences.

🚀 Skills

Programming: Python, C/C++, Javascript, Git, Scripting (Bash), LaTeX, HTML, Vim
Operating System: Linux, Windows
Technologies: Computer Vision, Machine Learning, Deep Learning, Natural Language Processing, Cloud Computing, MLOPS
Frameworks: Tensorflow, Pytorch, Keras, OpenCV, Fastapi, Numpy, Pandas, Matplotlib, Seaborn, Selenium, Hugging Face

📈 Skill Assesments

Machine Learning, in top 30% among 188K participants.
Python (Programming Language), in top 30% among 1 million participants
Problem Solving (Basic), HackerRank Skill Certifications
Python (Basic), HackerRank Skill Certifications

🌟 Achievements

CTO Appreciation Award, CTO Appreciation Award Giving Ceremony, SSL Wireless, Dec 2022.
High Flyer Hero, Mid Year Performance Award, SSL Wireless, Sep 2022.
Winner, Idea Contest, Esho Robot Banai, Aug 2017.
Runner-Up, Soccerbot, NSU Bit Arena, Jan 2019

📚 Publications

“Fabrication of Smart Eye Controlled Wheelchair for Disabled Person”, Md. Anisur Rahman, Md. Abdur Rahman, Md. Imteaz Ahmed, and Md. Iftekher Hossain, International Conference on Big Data, IoT and Machine Learning (BIM 2021)
“A Deep Learning Approach to Count people Using Facenet Architecture”, Md. Iftekher Hossain, Md. Sakirul Alam, International Conference on Emerging Trends in Industry 4.0 (2021 ETI 4.0), Raigarh, Chhattisgarh, India
“An Efficient Way to Recognize Faces Using Mean Embeddings”, Md. Iftekher Hossain, Sama‑E‑Shan and Homayun Kabir, IEEE First International Conference on Advances in Electrical, Computing, Communications and Sustainable Technologies (IEEE ICAECT 2021)
“An Approach to Maximize The Integrated Safety System For Two Wheelers”, Md. Abdullah Al Mamun, Bably Das, Sama‑E‑Shan, and Md. Iftekher Hossain | Conference Name 2
“Low Cost Deaf Communication System”, Md. Iftekher Hossain, Marzan Alam, and Md Rakibul Islam Prince | 1st National Conference On Energy Technology and Industrial Automation (NCETIA 2018)