Dallas-Fort Worth Metroplex
I'm a Staff Machine Learning Engineer, specializing in Natural Language Processing (NLP). My interest in NLP began when I studied abroad in Poland during my Bachelor's in Computer Science. The challenge of decoding language data (such as the difficult Polish language), through the use of technology, began to fascinate me. So, after finishing my Bachelor's in the US, I completed a Master's in Language Science and Technology in Germany, and started my career. Throughout my career, I have built: - multilingual search engines - question answering services - spam classifiers for social media - multilingual lemmatizers - and many other NLP solutions. I built these using deep learning, transformers, and other kinds of machine learning algorithms. Now, what excites me the most is the exponential degree at which these NLP-related AI/ML algorithms are advancing. Even with existing transformer-based models, the potential for creating unprecedented artificial intelligence is now immense. As Oliver Wendell Holmes Sr. once said, "Language is the blood of the soul into which thoughts run and out of which they grow." By building AI systems that excel at natural language processing, we extend our capabilities beyond the mere processing of text data. We create solutions capable of understanding ideas. Feel free to get in touch!
• Architected and scaled a RAG chatbot platform for 9,000+ users, achieving an 80% satisfaction rate and a 1.8x increase in instructor engagement, using LangGraph, Databricks, PostgreSQL, Redis, DynamoDB, and Docker. [Case Study: https://www.wgu.edu/blog/wgu-new-ai-tool-drove-more-instructor-calls2604.html ] • Established a high-velocity engineering culture, increasing team productivity by 200%, by transitioning the team to a Continuous Delivery (CD) model and enacting best practices in Git workflows and Agile Scrum methodology. • Implemented rigorous development standards by creating a full test suite, automating CI/CD pipelines with Github Actions, and establishing real-time error monitoring to ensure code quality and platform stability.
• Architected and scaled a core Retrieval-Augmented Generation (RAG) chatbot platform to production, successfully serving over 9,000 users with advanced AI capabilities, leveraging LangGraph, Databricks, PostgreSQL, Redis, and Docker. • Implemented rigorous development standards by creating a full test suite, automating CI/CD pipelines with GitHub Actions, and establishing real-time error monitoring to ensure code quality and platform stability. • Significantly improved platform access and user support by implementing Single Sign-On (SSO), integrating granular student access validation, and reducing help desk response time by 90% through established team strategy and comprehensive documentation.
• Boosted automation by 25% by performing feature engineering, training, and deploying new machine learning (ML) models for medical coding of Electronic Health Record (EHR) data. • Implemented data engineering Extract, Transform, Load (ETL) processes for maintaining freshness of data lake in Snowflake data warehouse, using SQL. • Leveraged advanced data analysis and clustering on unstructured NoSQL data to perform topic modeling and uncover key data science insights using NLTK, Scikit-Learn, and Pandas. • Translated Jupyter notebooks into Git versioned scripts on GitHub for improved model validation processes and A/B testing of models, and recorded documentation in Asana.
• Improved information extraction accuracy by 15% by fine-tuning large language models (LLMs) from Hugging Face Transformers hosted on Amazon Web Services (AWS). • Added new multilingual machine translation feature to core SaaS product offering for improved user experience, utilizing deep learning with Keras and PyTorch, natural language processing (NLP) methods with Gensim, and integrating with production code using Scala. • Reduced word splitting error rate with generative language models, trained using unsupervised learning on scraped web data, after peforming data preprocessing and data cleaning with SpaCy in Python. • Deployed deep learning models to cloud computing production environments in Docker containers with REST APIs. • Optimized computations on vector databases using NumPy, demonstrating analytical and problem solving expertise. • Demonstrated exceptional communication and teamwork in consulting regularly with cross-functional product teams in order to add additional features, within an Agile Scrum-based workflow, using Atlassian Jira and Confluence for effective project management, time management, and documentation.
• Boosted production text classification accuracy by 30%, by implementing deep learning models for spam filtering, sentiment analysis, and social media management. • Redesigned workflow for in-house annotators, for faster and more accurate labeling of training data for supervised learning. • Informed strategic decision-making by performing data analysis and topic modeling with Gensim, and presenting findings to stakeholders, utilizing data visualization with Matplotlib and Microsoft Excel for data reporting.
• Updated existing LSA-based question-answering engine in Java to a new convolutional neural network question-answering engine in Python, using Keras and Tensorflow. • Built and administered website using WordPress, CSS, and HTML.