Hello
I'm MD
Shahariar
Alom
Experienced machine learning engineer with proficiency in Python, TensorFlow, Keras, Scikit-learn, and other ML and DL frameworks. Skilled in end-to-end machine learning pipeline development, including data collection, preprocessing, model training, and evaluation. Developed deep learning models for image classification, object detection, Segmentation, face recognition, and NLP. Achieved high accuracy in SSD and other CNN algorithms. Seeking a role as a machine learning engineer to apply my AI development expertise to deliver innovative solutions.
Our Go-To ML Team: Delivering Custom Business Solutions for Every Challenge
PORTFOLIOS
My Experiences
Jan 2022 to Jan 2024 (2.7 year)
Zero Gravity Multimedia & IT Solutions
AI/ML Engineer, Data Analysis, Data collection, Project Management, Data Annotation Lead (Object Detection App Project-remote)
➢ Project: Arabic Learning App
Description: Android and iOS app for Arabic language learning using real-time object
detection.
➢ Role:
• Collected and annotated data from multiple datasets, analyzed them, and performed
critical operations to accurately detect objects.
• Trained and fine-tuned an object detection model on a custom dataset.
• Implemented real-time inference on mobile devices.
• Collaborated with Developers to deploy the model.
➢ Impact:
• Enhanced Tourist Experience: The Arabic learning app, with its real-time object
detection feature, significantly enhances the language learning experience for tourists.
By identifying and translating objects in their surroundings, it empowers travelers to
communicate effectively and immerse themselves in local culture.
Jan 2020 to December 2021 (1 year)
Quantigo AI
Data Annotation Lead (remote)
➢ Project: Vehicles and Number Plate Annotation
Description: Led a team responsible for annotating vehicle and license plate data.
➢ Role:
• Oversaw data collection and annotation efforts.
• Ensured consistent labeling and adherence to guidelines
➢ Impacts:
• Improved law enforcement and parking management by accurately identifying vehicles
and license plates.
• Enhanced security and safety through efficient monitoring.
Jan 2020 to December 2021 (1 year)
Upwork
Data Scraping, Data Analysis, Data Annotation, Image Classification, Object Detection, Intense Segmentation (freelance)
➢ Project 1: Apple Segmentation Model
Description: Developed an image segmentation model specifically for apple quality
and count.
➢ Role:
• Annotated diverse apple images for training data.
• Trained and fine-tuned a YOLO (You Only Look Once) model to segment apples.
➢ Impact:
Enhanced Production Tracking:
• The segmentation model accurately identifies and classifies apples based on quality
(e.g., Good, defects).
• Businesses can track production batches, ensuring consistent quality and minimizing
waste.
Early Detection of Affected Apples:
• By detecting defects early in the production process, the model prevents subpar
apples from reaching consumers.
• Improves customer satisfaction and reduces financial losses.
➢ Project 2: Interior Design Categorization via Classification
Description: Developed an AI-powered system to categorize interior design styles
based on visual features.
➢ Role:
• Collected diverse data through web scraping from different sources
• Curated a balanced dataset of interior design images representing various styles (e.g.,
modern, minimalist, traditional).
• Trained a classification model to predict design categories.
➢ Impact:
Personalized Recommendations:
• The classification model accurately assigned design styles to user preferences.
• Enabled personalized interior design suggestions for clients.
Efficient Decision-Making:
• Simplified the design process by automating style identification.
• Streamlined client consultations and project planning
Jan 2020 - May 2023 (4.5 years)
Fiverr
Data Annotation, Data Analysis, Research, Critical operations on datasets, Deep Learning App (freelance)
➢ Role:
• Annotated data for Computer Vision, NLP.
➢Complete 50+ Projects with clients satisfection!
➢ Impact:
• Improved existing model accuracy like mediapipe, and handled complex operations
on various datasets.
Relevant Skills
Libraries/Frameworks:
Extra Skills
Projects I've worked on
Scan and Learn Arabic is a mobile application that allows users to scan a previously identified image and know its meaning in Modern Standard Arabic and the Emirati dialect.
The application targets visitors to various venues in the Emirate of Abu Dhabi and aims to make learning Arabic accessible, engaging, educational and user friendly.
It is an initiative between the Abu Dhabi Arabic Language Centre, the Department of Community Development, and the Department of Municipalities and Transport, that is aimed at promoting the use of the Arabic language and ensuring its preservation among the community.
In this project, I developed a system to detect and extract text from vehicle number plates using EasyOCR, an open-source Optical Character Recognition (OCR) library. The system leverages computer vision techniques to locate number plates in images and then applies EasyOCR to accurately read and extract the text. This solution is particularly useful for applications such as automated toll collection, parking management, and law enforcement. By integrating EasyOCR with real-time object detection, the system ensures efficient and reliable number plate recognition, enhancing overall operational efficiency.
In this project, I developed a system to detect and map facial landmarks using Google MediaPipe Face Mesh. The solution estimates 468 3D face landmarks in real-time, even on mobile devices, using a single camera input. This project leverages machine learning to infer the 3D surface geometry of faces, enabling applications such as augmented reality (AR), facial recognition, and virtual try-ons. By integrating MediaPipe Face Mesh with real-time processing, the system provides accurate and efficient facial landmark detection, enhancing user experiences in various interactive applications.
C e r t i f i c a t i o n s
Machine learning under the Hire and Train program of EDGE
Institution: Bangladesh Computer Council(BCC), ICT Division
2023
Why Hire Me?
With over 3.5 years of hands-on experience as a Machine Learning Engineer, I bring a wealth of knowledge and practical skills to the table. Here’s why I’m the ideal candidate for your project:
Deep Learning Expertise:
- Extensive Experience: I have a strong background in developing and deploying deep learning models, particularly in computer vision and natural language processing.
- Advanced Techniques: Proficient in using state-of-the-art frameworks like TensorFlow, PyTorch, and YOLO for tasks such as image classification, object detection, and segmentation.
- Innovative Solutions: I have successfully implemented cutting-edge techniques to solve complex problems, ensuring high accuracy and efficiency.
Proven Industry Experience:
- Upwork Projects: Over the past 3.5 years, I have worked on diverse projects on Upwork, ranging from data collection and annotation to model training and deployment.
- Impactful Projects: My work includes developing an Arabic learning app using object detection, optimizing apple production through segmentation models, and providing interior design recommendations using classification models.
- Government Endorsements: I have led projects endorsed by government entities, such as vehicle tracking and monitoring systems, showcasing my ability to deliver high-impact solutions.
Technical Proficiency:
- Data Handling: Skilled in collecting, annotating, and preprocessing large datasets to ensure high-quality inputs for machine learning models.
- Model Training: Experienced in training and fine-tuning models on custom datasets, optimizing performance for specific use cases.
- Deployment: Proficient in deploying models on various platforms, ensuring seamless integration and real-time inference capabilities.
Collaborative Approach:
- Team Leadership: I have successfully led annotation teams, ensuring consistent and accurate data labeling.
- Cross-functional Collaboration: Worked closely with developers, UI/UX designers, and other stakeholders to deliver comprehensive solutions.
- Client Satisfaction: Committed to understanding client needs and delivering results that exceed expectations.
By hiring me, you gain a dedicated professional with a proven track record of delivering innovative and impactful machine learning solutions. Let’s work together to bring your project to life!
What I do
Data annotation to model deployment, I’ve got you covered.
️ Services I Provide:
Object Detection & Intense Segmentation, Object Tracking
- Intense & Semantic Segmentation
Framework Proficiency:
- TensorFlow (TF) 2.13, 2.14, 2.15, TFLite (I will fix bugs in all these versions)
Model Training & Conversion:
- EfficientDet-Lite, SSDMobileNet, TFLite-Model-Maker
TFLite Model Optimization:
- Metadata injection for seamless Android app integration
YOLO:
- Custom resolution object detection & segmentation
- Yolo Tiny
PyTorch Use Cases:
- Real-time applications
- Medical image analysis
- Natural language processing
Additional Expertise:
- NLP Chatbot Development
- Image Generation
- Traditional ML Models (Recommendation Systems, Predictions, etc.)
Bug Fixes & Installation Support Struggling with installation or training? Ill help you navigate through the complexities and get your models up and running.
Data Annotation
Our expert team is proficient in any type of annotation/Labeling and does any annotation related to machine learning and deep learning. Some of the Popular Formats Are
Modeling and Bug fixing
Developed robust machine learning models to solve complex problems and ensure high accuracy. Identified and resolved critical bugs, enhancing system stability and performance.
Deep learning powered Mobile app
If you are looking for a user-friendly app that will attract more mobile users, I can help you design and build a platform with the latest and trendiest look and feel.
I had the pleasure of working with Shahariar on developing an object detection Android app, and I couldn’t be more satisfied with the results. Shahariar demonstrated exceptional expertise in machine learning and computer vision, delivering a robust and efficient solution. The app accurately detects and identifies objects in real-time, significantly enhancing our operational efficiency. Shahariar’s attention to detail, responsiveness, and commitment to quality made the entire process seamless. I highly recommend Shahariar for any machine learning projects!
Manfred Prantle
Founder of 4smartmachine
Excelent Work I had the pleasure of collaborating with Shahariar on a YOLOv8 instance segmentation project.Shahariar really knows their stuff when it comes to deep learning. They picked the right model (YOLOv8) and nailed the training process. Great job, well done. He is very reliable and does his work in a way as to meet and exceed the buyers expectations. Even with a simple dataset, he achieved accurate results. Plus, they handled data-annotation part like a pro. I can highly recommend him.
Driss En-Nejjary
Ghent UNiversity
EDUCATIONAL QUALIFICATIONS
Bachelor of Science in Electrical and Telecommunication Engineering (B.sc in ETE)
University : Daffodil International University
Passing Year : 2022
Result : CGPA 2.85 Out of 4.00
Higher Secondary Certificate (H.S.C)
Institute : Holy Land College
Board : Dinajpur
Passing Year : 2016
Result : CGPA 3.75 Out of 5.00
School of Secondary Certificate (S.S.C)
Institute : Kherkati High School
Board : Dinajpur
Passing Year : 2014
Result : GPA 5.00 Out of 5.00
LANGUAGE PROFICIENCY
Language Reading Writing Listening Speaking
Bengali Excellent Excellent Excellent Excellent
English Excellent Good Good Good
PERSONAL INFORMATION
Full name : Md. Shahariar Alom
Father’s Name : Md. Moksedul Alom
Mother’s Name : Sajeda Khatun
Date of Birth : 12-01-1998
Marital Status : Unmarried
Religion : Islam
Present Address : Behind Maharaja School, Uttar Balubari, Dinajpur Sadar-5200, Dinajpur
Permanent Address : Behind Maharaja School, Uttar Balubari, Dinajpur Sadar-5200, Dinajpur
