Oct 2022 – Mar 2023 · 5 mos · Chicago, Illinois, United States

• Built autonomous LLM agents capable of executing multi-step analytics tasks, generating SQL, validating outputs, and generating automated summaries.
• Designed RAG+agent systems using vector DBs (FAISS/Pinecone/Chroma) for semantic search and real-time transit insights.
• Implemented LangGraph-based agent state machines for deterministic behavior and tool-governed actions.
• Built distributed ML & ETL workflows using Databricks, PySpark, ADF, Airflow, Dask, and Ray.
• Developed RLHF pipelines to refine chatbot and recommendation systems.
• Built cloud-native ML services using Docker, Kubernetes, and Azure ML for production-grade serving.

Feb 2021 – Aug 2022 · 1 yr 6 mos · Boston, Massachusetts, United States

• Built ML pipelines for forecasting, emissions modeling, and sustainability analytics.
• Applied NLP (TF-IDF, Word2Vec, NLTK, GPT-3) for document classification & automated summarization.
• Implemented CV models using TensorFlow/OpenCV for document and image classification.
• Created Power BI dashboards integrating predictive models with operational insights.
• Developed anomaly detection, ETL pipelines, and cloud-native deployments using AWS SageMaker.

Jun 2017 – Dec 2020 · 3 yrs 6 mos · Jersey City, New Jersey, United States

• Developed and deployed supervised ML models (logistic regression, decision trees, XGBoost) for customer churn prediction and fraud detection using Python (Scikit-learn, XGBoost).
• Built scalable, cloud-native ETL and ML pipelines on Google Cloud (BigQuery, Dataflow, Pub/Sub, Dataproc) integrating real-time data into predictive models.
• Implemented NLP pipelines for text classification and customer query segmentation using NLTK, SpaCy, and TF-IDF, improving call center routing accuracy by 25%.
• Designed and automated MLOps workflows with Apache Airflow, enabling CI/CD for data pipelines and retraining ML models weekly.
• Built feature stores and implemented feature engineering best practices across financial and customer datasets.
• Containerized ML services using Docker and deployed on Kubernetes, enabling scalable fraud scoring APIs.
• Created model monitoring dashboards using Prometheus and Grafana to track model drift and prediction performance.
• Collaborated with data scientists to operationalize experimental models, ensuring reproducibility and version control using MLflow and Git.
• Applied explainable AI techniques (SHAP, LIME) to communicate risk model outputs with business stakeholders and compliance teams.
• Translated business KPIs into SQL queries, curated datasets, and automated dashboards.
• Implemented data quality checks to resolve data gaps, duplicates, and inconsistencies.
• Built advanced dashboards in Tableau and Looker, integrating live predictions from ML APIs to enhance executive decision-making.

Apr 2013 – May 2017 · 4 yrs 1 mo · Irving, Texas, United States

• Analyzed business processes across engineering and operational teams to document system requirements and workflow needs.
• Created BRDs, FRDs, process diagrams, and reporting specifications for ETL, analytics, and maintenance systems.
• Participated in UAT cycles including test planning, test execution, defect logging, and system validation.
• Documented AS-IS/TO-BE workflows to support process improvements and automation initiatives.
• Collaborated with cross-functional teams to translate business requirements into technical enhancements.
• Supported development and validation of dashboards for operational reporting and maintenance insights.
• Conducted data analysis, validation, and rule documentation for multi-source operational datasets.
• Worked with system owners to enhance ETL workflows and improve data accessibility and accuracy.
• Contributed to system modernization through documentation, requirement translation, and workflow mapping.
• Supported production issue triage, defect analysis, and continuous improvement efforts.