# Navya K.

> Machine Learning Engineer/AI | Specializing in Reinforcement Learning | MS in Computer Science

Profile: https://flows.cv/navyak

I’m an AI/ML & Data Science professional with 7 years of experience building end-to-end analytics and production-grade ML systems across Retail, Media, Telecom, E-commerce, Banking, and Healthcare. My work spans NLP, Generative AI, and Reinforcement Learning—most recently applying Actor-Critic methods to Named Entity Recognition (NER) to improve F1 and reduce false positives/negatives on real-world data.
I design and ship scalable solutions: Python/SQL/PySpark for data wrangling, TensorFlow/PyTorch/Keras for modeling, and FastAPI/MLflow for MLOps. I’ve built LLM apps with GPT-4o, Copilot, Gemini, LLaMA, using RAG (LangChain + Pinecone/Chroma/FAISS) to deliver accurate, grounded answers. On the data side, I leverage Hadoop, and Airflow with Snowflake/Databricks/BigQuery to power enterprise analytics.

I’m passionate about aligning AI with business goals—accelerating decisions, automating workflows, and creating measurable value—while mentoring teams and collaborating in Agile environments.

## Work Experience
### Senior NLP Data Scientist @ UnitedHealth Group
Jan 2023 | United States
Designed ML pipelines for automated classification & summarization of unstructured healthcare documents using Scikit-learn, TensorFlow, and spaCy, cutting manual review time.
Built predictive models (Logistic Regression, Random Forest) to identify high-priority patient cases, improving clinical risk assessment workflows.
Implemented NER and POS tagging with spaCy to extract clinical terms & PHI for compliance and analytics.
Led migration to LLM-powered clinical intelligence using GPT-4, LangChain, RAG, enabling advanced summarization & semantic search.
Integrated FAISS and vector databases for rapid, context-aware clinical information retrieval.
Fine-tuned BioBERT & ClinicalBERT to boost accuracy on domain-specific text tasks.
Built real-time document ingestion workflows with PySpark & Apache Airflow for automated scheduling and scaling.
Delivered outputs via Tableau dashboards and FastAPI APIs, integrating with hospital systems.
Containerized ML/LLM workflows with Docker, deploying at scale on AWS SageMaker & Azure ML Studio.
Established data validation & QA pipelines with Pandas and Great Expectations, ensuring high data integrity.

### Machine Learning Engineer @ M&T Bank
Jan 2021 – Jan 2023
Developed advanced machine learning models to optimize the loan approval system, leveraging customer demographics, credit scores, and financial history to improve predictive accuracy and reduce default rates.
Designed scalable ML pipelines in Python and validated statistical assumptions in R, processing high-volume transactional data using Scala and PySpark on Hadoop-based distributed systems.
Performed advanced feature engineering and dimensionality reduction using Pandas, NumPy, pandas, and scikit-learn for large-scale datasets, improving model training efficiency and accuracy.
Applied ensemble learning techniques including XGBoost to improve classification accuracy and model robustness for imbalanced datasets.
Experimented with deep learning models using TensorFlow and Keras for customer churn prediction and unstructured document classification tasks.
Built OCR-based ingestion workflows with OpenCV, integrated with spaCy and NLTK for text extraction, named entity recognition (NER), and automated document classification.
Designed and maintained interactive dashboards in Tableau and Power BI to visualize KPIs for credit risk, churn rates, and product performance, enabling faster executive decisions.
Created model diagnostics visualizations (confusion matrices, ROC curves, feature importance plots) using Matplotlib and Seaborn for interpretability and model validation.
Developed real-time and batch ETL pipelines using Kafka and Apache Airflow to integrate data from multiple banking systems into centralized repositories.
Deployed ML solutions on AWS and explored deployment alternatives on Azure ML Studio for flexibility and scalability.
Managed hybrid data environments with PostgreSQL, MySQL and MongoDB, to meet diverse storage and retrieval needs.
Implemented CI/CD pipelines with Jenkins for automated testing and deployment of containerized ML applications using Docker.
Used GitHub for source control, enabling collaborative model development.

### Data Scientist @ Kroger
Jan 2018 – Jan 2021
My role involved transforming data processing and pricing strategies to enhance operational efficiency and customer engagement.  

• Developed automated data ingestion workflows, reducing manual intervention by 80% and improving pipeline resilience.  
• Conducted market elasticity studies, integrating competitor trends and real-time demand signals into pricing models.  
• Collaborated with marketing to launch targeted campaigns, resulting in increased revenue per transaction and customer satisfaction.


## Education
### Master of Science - MS
Texas Tech University

### PGP
Great Learning

### Bachelor of Technology - BTech
Jawaharlal Nehru Technological University Kakinada (JNTUK)


## Contact & Social
- LinkedIn: https://linkedin.com/in/navya-kamireddy
- Email: mailto:knreddy507@gmail.com

---
Source: https://flows.cv/navyak
JSON Resume: https://flows.cv/navyak/resume.json
Last updated: 2026-04-16