Generative AI & LLMs:
I am a Machine Learning Engineer with a passion for delivering AI-driven solutions that create measurable impact. Specializing in designing and deploying end-to-end machine learning pipelines, I leverage cloud platforms like AWS and advanced frameworks to craft scalable, innovative solutions. With experience across industries such as logistics, healthcare, and e-commerce, I focus on optimizing processes, reducing costs, and enabling data-driven decision-making. My work combines technical expertise with a strong understanding of business objectives, ensuring meaningful outcomes for organizations.
Core Areas of Expertise
End-to-End Machine Learning Pipelines: Proficient in data ingestion, cleaning, feature engineering, model development, and deployment, managing datasets of 50,000+ rows.
Advanced Machine Learning Models: Expertise in Random Forest, XGBoost, Deep Neural Networks, and Generative AI models, achieving performance improvements up to 20%.
Generative AI & NLP: Leveraging GPT models, Hugging Face, and OpenAI to automate content creation and deliver predictive insights.
Model Optimization: Skilled in GridSearchCV, RandomizedSearchCV, and Optuna, enhancing model accuracy and efficiency by up to 30%.
MLOps & CI/CD Pipelines: Developing robust, reproducible workflows using Docker, Kubernetes, MLflow, and Streamlit, reducing deployment time by 40%.
With dual master’s degrees in Business Information Systems and Software Engineering, I bring a unique blend of technical and business acumen to every project. My goal is to deliver AI-driven innovations that not only solve complex problems but also create measurable business value.
Published a research paper ”Chatbot for Disease Prediction and Treatment Recommendation using
Machine Learning”, High Technology Letters International Journal, ISSN NO: 1006-6748, Volume 27, Issue
6, June 2021, pp. 354-358. → Access the link here
Worked as a Graduate Research Assistant under Associate Professor Donghee Wohn, researching the Roblox metaverse and its influence on online dating dynamics among Gen Z. Contributed to a 30% increase in research productivity.
Developed and deployed an end-to-end machine learning project for real-time logistics company for truck delay prediction by leveraging AWS RDS, SageMaker, Hopsworks feature store and model registry, MLflow, and Streamlit, integrating advanced feature engineering, hyperparameter tuning, and model evaluation. Delivered a 15% accuracy improvement, 40% reduction in deployment time, and $200,000 annual cost savings through optimized routing and enhanced operational efficiency.
Source Code --> Github
Integrated AI into a user-facing application by developing an end-to-end ML workflow to predict house prices using regression models such as Linear Regression, Random Forest, leveraging Python (Pandas, NumPy) for data cleaning, feature engineering, and outlier detection. Fine-tuned models with GridSearchCV, improving accuracy by 15%, and logged parameters and model performance using MLflow for reproducibility and model tracking. Deployed the solution using Docker and Kubernetes, reducing deployment time by 30%. Improved model accuracy by 15% through fine-tuning with GridSearchCV. The final model reduced pricing errors by 20%, enhancing decision-making and customer satisfaction.
Source Code --> Github
Developed a medical chatbot for preliminary disease diagnosis and treatment recommendations using machine learning techniques (Naive Bayes, Decision Tree) and natural language processing (NLP). The chatbot interprets user-provided symptoms to predict possible diseases and suggests treatments, reducing the need for hospital visits for minor ailments. Achieved 92% accuracy in diagnosing ailments, improving healthcare accessibility and efficiency by 25%.
Source Code --> Github
Language Translator is an AI-powered tool for text and voice translation using Hugging Face's MarianMT, Llama 3.2 via the Ollama API, and LibreTranslate. It supports multi-language translation with text-to-speech and speech-to-text functionality.
SourceCode --> Github