Hi, Welcome to my website!
I’m Saidur Rahman Sujon
Full-Stack Developer | ML/AI Engineer
About Me
As a B.Sc. in Computer Science and Engineering (CSE) graduate, my professional journey has been focused by the roles as a Full-Stack Developer & ML/AI Engineer. My expertise lies in building full-stack web applications. Also integrating latest AI technologies such as Generative AI with existing frameworks. Being also proficient in Data Science and crafting efficient ML algorithms for extracting actionable insights from data. My portfolio spans diverse projects, including e-commerce apps and immersive 3D games. My research interest area lies in GenAI, NLP, CV and Data Science.
What I Do
I develop AI models & pipelines for industry level solution. I also design systems & build softwares from Front-end to Back-end, especially MERN stack applications, with optimized code writing.
Research and Development (R&D)
Design and Development (D&D)
Development
Development
My Skills
I am a fast learner, specialized in multitude of skills required for Software Development and Machine Learning operations. From design to development, research to deployment, my strenght in multiple programming stack allows me to leverage the ability to build innovative solutions.
Portfolio
Takaramichi
Duration: 2024.9-2024.10 (1 month)
Takaramichi is a web application designed to streamline the management of inbound accommodation and ski rentals in the popular Japanese resorts of Nozawa Onsen and Hakuba Village. This user-friendly platform simplifies booking, rental management, and customer interactions, tailored specifically for businesses in these regions. Built with React js and Tailwind CSS.
- React Js
- HTML5
- Tailwind CSS
- Node Js
- MongoDB
LlamaChirp: Chat with
PDF using Local LLMs
Duration: 2024.4-2024.5 (1 month)
LlamaChirp is an innovative project enabling users to engage in dynamic conversations with PDFs to extract and comprehend information using locally hosted LLM variants of Ollama by integrating Retrieval-Augmented Generation (RAG) techniques. Experience seamless interaction, precise responses, and personalized conversational experiences with LlamaChirp.
- Chainlit
- LangChain
- Ollama
- ChromaDB
- RAG
- Gen-AI
- LLM
- Python
- PyPDF2
Chatter AI: Chat with
Website
Duration: 2024.3-2024.4 (1 month)
Chatter.AI is a web application that allows you to chat with websites and extract information from them. It uses the OpenAI API and LangChain library to process and understand the content of the website and allow users to interact with it on the Streamlit interface. It has two output formats: JSON and Q/A, including I/O configuration flexibility.
- Streamlit
- LangChain
- OpenAI
- ChromaDB
- Gen-AI
- LLM
- Python
- BeautifulSoup
- RAG
Job Lagbe
Duration: 2024.3-2024.6 (3 months)
A job hunting Full-Stack MERN web application that helps candidates find their dream jobs and job recruiters find their best candidates. It is a dynamic React front-end application created using VITE and contains pages: Landing, Error, Register, and Dashboard with seamless page transitions. Established a robust server application using Node.js, Express, MongoDB, including navigating with React Router 6.
- MongoDB
- Express.js
- React.js
- Node.js
- HTML5
- CSS3
RMA agro
Duration: 2024.1-2024.2 (1 month)
A simple website for an agricultural company named RMA agro, built with HTML5, CSS3 and JavaScript framework. Users will be able to order directly from farmer and enjoy fresh and delicious foods. The website is fully responsive and the website is hosted on GitHub Pages.
- HTML5
- CSS3
- JavaScript
ML: Stock Market Price
Prediction
Duration: 2023.11-2023.12 (1 month)
Unveiling stock market trends thorugh news report analysis and price prediction using numerical and textual data. Machine Learning methods like Feature Engineering, Data Preprocessing, Modeling, Evaluation and Plotting etc. were implemented in this project. The final output shows the predicted trends with the testing close price with greater accuracy.
- Machine Learning
- NLP
- NumPy
- Pandas
- Seaborn
- Matplotlib
- Keras
- NLTK
EDA: US-based Retail
Store
Duration: 2023.11-2023.12 (1 month)
Exploratory Data Analysis of a US-based reatil store highlights the sales report with interactive data visualization graphs and charts in PowerBI. This project was an internship task of The Sparks Foundation. Slicers, geographical analysis, matrix, hierarchy, tree map, bar, gauge and pie charts etc. were used to finding valuable insights from the data.
- Data Analytics
- Data Visualization
- PowerBI
- Excel
- Data Analysis
Rubiks Realm
Duration: 2023.6-2023.8 (3 months)
Fully interactive and customizable 3D Rubik's Cube, with options for changing the cube size, flip-type, scramble length, camera angle, and color scheme. keeps track of user statistics, options for starting a new game, resetting the cube, and customizable game settings
- HTML5
- CSS3
- JavaScript
- Three.js
BharaDibo
Duration: 2023.2-2023.4 (3 months)
BharaDibo is an ASP.NET MVC 5 web app for renting individual properties, enforcing policy constraints. Admins manage user accounts, product info, and customize features using PHP and MS SQL. It is a valuable tool for efficient rental property management and product discovery for both renters and landlords.
- ASP.NET MVC 5
- HTML5
- CSS3
- SCSS
- JavaScript
Weekend Music Store
Duration: 2022.6-2022.8 (3 months)
An e-commerce management system for online shopping, created with JavaFX, SceneBuilder on Netbeans; implemented with RDBMS. Multiple dashboards designed for users and admins, Inventory management, Sales Analysis, Product searching, and credit card transaction gateway are implemented. The project is designed to be used for both online and offline shopping as self-service.
- JavaFX
- SceneBuilder
- Netbeans
- MySQL
- CSS
Arithmetic Logic Unit Implementation
Duration: 2022.3-2022.6 (3 months)
This project involves building an Arithmetic Logic Unit (ALU) for a Digital System Design lab. The ALU performs binary operations such as addition, carry addition, complement, and AND logic. Essential components include a power adapter, logic board, wires, LED bulbs, and switches. It demonstrates practical implementation of digital logic concepts. The project showcases both hardware integration and functional logic design.
- Logic Board
- Power Adapter
- Wires
- LED bulbs
- Switch
Fullmetal Alchemist
Redemption
Duration: 2021.6-2021.8 (3 months)
A 2D Shooting game inspired by the anime Fullmetal, made with igraphics engine using OpenGL library in C and C++. Intuitive keyboard-controlled aiming and movement, firing bullets towards ground and sky enemy fleet, that add a unique challenge, multiple levels, each with different enemy adversaries, health points availability, immersive soundtracks and visuals, easy controls and score board.
- C#
- C++
- iGraphics
- OpenGL
- MS Visual Studio
Experience
Data Science and Business Analytics Intern
(Nov 2023-Dec 2023)
As an Intern, I was assigned with different real world
projects to extract insights from the data and build ML
models for effecient prediction.
Key responsibilities: Performing data
acquisition, cleaning, and preparation using NumPy and
Pandas for exploratory data analysis. Applying statistical
methods and machine learning algorithms for trend
prediction and classification. Creating visual
representations with PowerBI to effectively communicate
findings.
Full Stack Development
I build personalized websites, seamlessly integrating Front-end and Back-end technologies. Proficient in HTML, CSS, JavaScript, React.js, and Vue.js, I execute robust server-side logic with Node.js and Python, managing databases (MongoDB, MySQL, PostgreSQL) for efficient data handling. I ensure smooth integration, API development, and collaborative version control using Git. My optimized solutions exceeds client expectations.
Education
B.Sc. in Computer Science and Engineering (CSE)
(2019-2023)
CGPA: 3.578 / 4.00
Higher Secondary School Certificate (HSC)
(2016-2018)
GPA: 5.00 / 5.00
Professional Exam Certifications
ITEE Level-II: Fundamental Exam (FE)
Full-Passer (FE01-0172) – April-2024
Half-Passer (FE01-0160) – October-2023
Topics Covered: Programming, Logic, Data Structures (Lists/Arrays, Queues/ Stacks, Trees/Graphs), Sorting/Searching, OOP, DevOps, Basic Linux, and Cyber Security.
Professional Training Certifications
Bangladesh-Japan ICT Engineers Training Program (B-JET):
Basic Plus Course
University of Miyazaki | Japan
October 21, 2024 - December 6, 2024
Skills gained: N4 to N3 (partial) level Japanese Language Proficiency, learned more about Japanese Traditions and Business Ethics, Rules and Regulations, IT skills including Design Pattern Techniques, Cloud computing, CI/CD pipelines. Created video contents on various topics in Japanese Language introducing Bangladeshi culture, Developed Nihongo project with collaborating with Takachiho High School students of Miyazaki Prefecture.
Bangladesh-Japan ICT Engineers Training Program (B-JET):
Basic Course
University of Miyazaki | Japan
June 3, 2024 - October 18, 2024
Skills gained: N5 to N4 level Japanese Language Proficiency, Japanese Culture and Business Manners, IT skills including Project Management techniques, Github collaboration, Developed a Full-Stack Online Learning Management System with Admin, Teacher and Student roles built with Next.js, PostgreSQL and Microservice Architecture. Participated in Programming Contest for B-JET 14th batch held by BJIT and B-JET authority.
Skills gained: Prompt engineering, Utilizing Large Language Models (LLMs) like ChatGPT, Crafting prompt patterns, Developing complex applications to solve particular problems, Using tactics to solve mathematical problems and generate desired results from given data. Zero-shot, Few-shot tactics, structured generative format etc. were utilized.
Skills gained: Fundamentals of Python, using databases with Python, accessing web data, data structures, DBMS, parsing JSON and XML using Python. Capstone: retrieving, processing, and visualizing data. These skills are gained from the Python specialization course, enabling comprehensive data management and analytical capabilities in real-world applications.
Skills gained: Understanding various encryption algorithms and techniques, their benefits, limitations, Knowledge of various authentication systems and types, Differentiating between authentication and authorization, Evaluating potential risks and recommending ways to reduce them, Assisting others in grasping security concepts and protecting themselves.
Skills gained: Ability to describe computer networks using a five-layer model, understand all of the standard protocols involved with TCP/IP communication, employ effective network troubleshooting tools, manage network services like DNS and DHCP, and grasp cloud computing concepts including various service models and cloud storage solutions.
Skills gained: Foundational skills in Information Technology, including understanding the binary system, assembling computers from scratch, and installing operating systems, problem-solving methodologies and soft skills essential for the IT field, such as communication and teamwork.
Research & Publications
Published: 14 October 2024 | Version 1
DOI:10.17632/pdtyb7df6z.1
Our dataset, BSMDD, was collected from various open social media platforms and translated and annotated by native Bengali speakers with expertise in both language and mental health. It contains 21,910 cleaned samples, including 10,961 labeled as Depressed and 10,949 as Non-Depressed. The dataset is publicly accessible, providing a valuable resource for further research in depression detection in Bengali social media content. The expert annotation process, conducted by professionals, ensures high validity, making BSMDD particularly important for advancing mental health research through social media analysis.
- Mental Health
- Data Science
- Bengali NLP
- Social Media Data
- GenAI
- Explainable-AI
- Psychology
Harnessing Large Language Models Over Transformer Models for Detecting Bengali Depressive Social Media Text: A Comprehensive Study
Publication: Volume 7, June 2024, 100075
This study focuses on early detection of depression, particularly in extroverted social media users, using LLMs such as GPT 3.5, GPT 4 and our proposed GPT 3.5 fine-tuned model DepGPT, as well as advanced Deep learning models(LSTM, Bi-LSTM, GRU, BiGRU) and Transformer models(BERT, BanglaBERT, SahajBERT, BanglaBERT-Base). The study categorized Reddit and X datasets into "Depressive" and "Non-Depressive" segments, translated into Bengali by native speakers with expertise in mental health, resulting in the creation of the Bengali Social Media Depressive Dataset (BSMDD).
- LLMs
- Transformers
- Deep Learning
- Prompt Engineering
- Fine-tuning
- Explainable-AI
- Psychology
- GenAI
Contact Me
If you like my work and want to avail my services then drop me a message using the contact form.
Or get in touch using my email, skype or WhatsApp number.
See you!