Senior Software Engineer at Pace Harmon

Developed predictive and recommendation models using Python, PyTorch, and scikit-learn for ranking, classification, and behavioral analysis. Improved decision quality by ~25% via feature engineering and iterative experimentation. Conducted EDA, statistical testing, feature selection, and error analysis across structured and unstructured datasets, and translated findings into product/business recommendations. Built NLP workflows for text classification, entity extraction, semantic matching, and summarization; applied LLMs and RAG for selected unstructured-data problems with retrieval evaluation and guardrails. Implemented scalable data/feature pipelines with Pandas, PySpark, Airflow, and Kafka; designed analytical data layers using PostgreSQL, Snowflake, Redis, and vector databases. Managed experiments and model lifecycle with MLflow; deployed training/scoring workloads across AWS, Azure, and GCP using Docker and Kubernetes. Added automated validation and production model monitoring, an

Senior Software Engineer at Access Global Group

Built and operated distributed ML training workflows using PyTorch, TensorFlow, Databricks, PySpark, and MLflow, improving training throughput by ~30%. Created reusable training infrastructure for experiment tracking, hyperparameter tuning, checkpointing, model comparison, and reproducible runs. Implemented feature and training pipelines with Airflow, Spark, Snowflake, dbt, and Pandas, including automated data preparation, feature generation, and scheduled retraining. Developed NLP and deep-learning models with Hugging Face and PyTorch for text classification, semantic similarity, and information extraction; established evaluation using cross-validation and task metrics such as precision, recall, F1, and ROC-AUC with error analysis and baseline comparisons. Productionized training and inference with AWS SageMaker, Docker, Kubernetes, and MLflow Model Registry, supporting batch scoring, versioning, and rollback. Added CI/CD and ML monitoring using GitHub Actions, Prometheus, Grafana, an

Software Engineer at VMware

Developed ML models in Python, scikit-learn, and TensorFlow for classification, recommendation, anomaly detection, and predictive use cases. Prepared datasets with Pandas, NumPy, and SQL; performed cleaning, exploratory analysis, and feature engineering. Evaluated models using validation datasets and metrics including confusion matrices, precision/recall, and F1 with error analysis. Built repeatable preprocessing and feature pipelines for training and inference, and integrated model predictions into production applications via Python/Flask REST APIs. Deployed containerized ML workloads across AWS and Azure, improving deployment efficiency by ~25%. Collaborated with data, engineering, and product teams on model integration, validation, and production troubleshooting.