Machine Learning Fort Wayne

Rifat Jahan Rezwan

Fort Wayne, IN - ***** Phone: 260-***-**** Linkedin Email Github Kaggle Huggingface EDUCATION

Master of Science Computer Science Purdue University Fort Wayne, IN, USA (Dec 24)

(Coursework: Data Structure, Machine Learning, Natural Language Processing, Expert System, Deep Learning, Data Analytics & Visual, Analysis and design of Algorithms, Software Engineering, Software Architecture, Human Computer Interaction, Database Systems, UX Design, Software Project Management, Web development, Cloud Computing Technology, Data Engineering) WORK EXPERIENCE

Software Development Intern MIE (Medical Informatics Engineering) Fort Wayne, Indiana, USA (Feb’24- Aug’24)

• Developed an AI-powered document management and retrieval system using Python, FastAPI, and LangChain, leveraging Ollama embeddings and NVIDIA LLM for context-aware query handling. Implemented FAISS for efficient vector search and retrieval, integrated Markdown processing, and containerized the application using Docker for scalable deployment. UI implemented with Gradio. Enhanced customer experience by optimizing document retrieval processes, leading to faster and more accurate query responses.

• Led the full-stack design, development, and Docker deployment of the system, focusing on architecture, document processing pipelines, and CI/CD automation with GitHub Actions. Optimized performance through efficient data handling and vector search strategies, ensuring accurate and fast document retrieval (RAG). Deployed machine learning models using Docker and Kubernetes, ensuring scalable and reliable production pipelines. Collaborated with cross-functional teams, including product managers and engineers, to align ML projects with business needs. Developed data pipelines and orchestrated workflows with Airflow for automated ML model training and deployment. Monitored deployed ML models using custom alerting mechanisms to handle model drift and edge- case scenarios effectively. Leveraged cloud services such as AWS and Azure for scalable model training and deployment, optimizing performance and cost.

Data Analyst ML Engineer ColorC Tech Co. Dhaka, Bangladesh (Jan’21 – Oct’22)

• Led advanced data analysis initiatives using Python to extract trends and actionable insights, significantly contributing to the achievement of key business targets. Utilized Python libraries such as Pandas for data manipulation and integrated Power BI and Looker for interactive dashboards and compelling visual reports, leading to a 15% increase in operational efficiency. Automated data pipelines and enhanced ETL processes using Python scripts, which improved data accuracy by 20% and accessibility for ongoing analysis. Full Stack Developer (Web) Amar shop Twenty-four Ltd. Dhaka, Bangladesh (Jan’19 – Dec’20)

• Led the development of the Amarshop24 e-commerce platform with React.js, enhancing customer engagement by 20% through improved UI/UX. Managed server-side operations using Node.js and Express.js, which improved API response times by 25%. Implemented responsive design principles, ensuring compatibility across various devices, contributing to a 30% increase in mobile user traffic.

Frontend Developer European Telecom Dhaka, Bangladesh (Oct’17 – Dec’18)

• Redesigned web interfaces with HTML5 and CSS3, improving website load time by 30%. Integrated AngularJS into the development stack, accelerating project delivery by 40%. Collaborated with UI/UX designers to translate designs into high-fidelity, responsive web pages, resulting in a 35% increase in user retention. SKILLS

Languages: Java, C#, Python, JavaScript, C++, R

APIs & Frameworks: RESTful API Services, Flask, Node.js, Rect Native, Express.js, FAST API, Streamlit, Gradio

TechStack : HTML5, CSS3, JavaScript, React.js, Express,, Langchain, Llamaindex, RAG, GraphRAG, Numpy, LangGraph, LangSmith, Tesnorflow, Pytorch

Cloud Technologies: GCP, AWS, AZURE

ML/DL: OpenCV, TensorFlow, Keras, Anaconda, Sci-kit-learn, NLTK, spaCy, NLP, LoRa, QLoRa, PEFT, LSTM, MLP, CNN, Neural Networks, CNN, MTCNN, Semantic Analysis

Data Analysis Libraries: Matplotlib, Seaborn, Plotly, Looker Version Control and Collaboration: Git, GitHub, Kaggle, Colab, Bitbucket, S3 Bucket, Athena, Lambda, Airflow, Kubeflow,MLflow Database Systems: MySQL, MongoDB, SQLite, Firebase, FAISS, PostgreSQL, NoSQL, DynamoDB, RDS, Neo4j, Qdrant, Chroma Distributed systems: Pycharm, Jupyter Notebook, Google Colab, VSCode, Tableau, Figma, WordPress, WIX, Webflow, Transformers, Pytorch, OpenAi, Nividia Nemo, Nim, Unsloth, B&W

Soft Skills: Problem-Solving, Creativity, Critical thinking, Teamwork, JIRA management.

Deployment: Docker, Kubernetes, CI/CD pipeline, GHCR, Jenkins, AWS (EC2, VPC, IAM, S3, SageMaker, ETL, Bedrock,Glue,Cloud Font) PROJECTS & RESEARCH

Sarcasm Detection on Reddit (Natural Language Processing)

• Implemented a sophisticated NLP model to detect sarcasm in Reddit comments, enhancing content understanding and sentiment analysis. Tech: Python (NLTK, spaCy, scikit-learn) ML Model: Logistic Regression, SPM, Random Forest, Decision Tree, Naïve Bayes, MLP Classifier

Automated Attendance System with Face Recognition (Deep Learning)

• Developed a cutting-edge system for automating attendance using facial recognition technology to streamline administrative processes. Tech: Face Detection with OpenCV (Python), Tensor Flow, Keras ML Model: Facenet, MTCNN VIS-CO: A Comparative Study for Intelligent Dashboard Visualization

• Developed an intelligent dashboard that enhances the visualization selection process, providing industry-specific insights with improved accuracy and efficiency. Conducted comparative analysis of manual and automated visualization techniques, identifying areas for improvement in existing tools. Experienced in deploying and refining large-scale ML models in cloud environments, ensuring high reliability and scalability.

• Datasets: Health insurance data from healthcare.gov and financial data from Kaggle. Viz Tools: Looker Studio, Power BI, custom Python. Model Train: Developed and trained the BERT model and Finetuned Meta/llama2 model for visualization prediction.

Contact this candidate