Mahesh babu kambala

AI/ML Engineer @ American Express

About

I’m Mahesh Babu Kambala — an AI/ML Engineer with 4+ years of experience designing and deploying scalable, production-grade AI systems across financial services and healthcare domains. My expertise lies in building end-to-end machine learning pipelines, fraud detection systems, and Generative AI platforms that operate reliably in real-world, high-volume environments. At American Express, I engineered enterprise-scale risk and fraud intelligence systems by developing Graph Neural Network (GNN)-based models to detect complex transactional relationships, reducing fraud losses by 37%. I enhanced credit risk modeling frameworks (PD, LGD, EAD) using ensemble learning techniques and built time-series forecasting systems to support proactive risk monitoring and stress testing. I’ve also implemented Retrieval-Augmented Generation (RAG) pipelines using GPT and LangChain to automate regulatory documentation workflows, reducing reporting preparation time by 55%. My work integrates MLOps best practices using MLflow and Kubeflow, along with explainable AI tools (SHAP, LIME) to ensure model transparency and regulatory compliance. Previously at HCL, I developed predictive healthcare analytics systems during COVID-19, including survival analysis models, patient readmission risk prediction (improving accuracy by 28%), and NLP pipelines for extracting insights from clinical notes using spaCy and healthcare standards like ICD-10 and SNOMED. Technically, I specialize in: • Machine Learning & Deep Learning • Graph Neural Networks • Generative AI & LLM Systems • Retrieval-Augmented Generation (RAG) • Azure Databricks & Spark • Python, SQL, Distributed Computing • MLOps (MLflow, Kubeflow) • Explainable AI & Model Governance I’m passionate about building AI systems that are not only accurate, but scalable, explainable, and production-ready. Always open to collaborating on advanced AI platforms, scalable ML infrastructure, and impactful real-world intelligent systems.

Country

United States

City

New York

Industry

Information Technology & Services

Skill

Server Side, Software Construction, Mean Stack, Java Development, Server Side Programming, Angular, Architecture, Jira, Kotlin, Spring Boot, Microsoft SQL Server, Amazon Web Services (AWS), Debugging, TypeScript, Docker, Flutter, HTML, SQL, Java, AngularJS

Experience

AI/ML Engineer

American Express

2023-8 - Present · 3 yrs 2 mos

Albany, NY

Project: AI-Powered Risk & Fraud Intelligence Platform (ARFIP)  Engineered scalable data pipelines using Airflow, dbt, and Spark on Azure Databricks to ingest and validate high-volume financial data from Bloomberg, FactSet, and Open Banking APIs, ensuring reliable data for risk and fraud analytics.  Conducted in-depth exploratory data analysis and advanced feature engineering on large multi-source financial datasets to identify fraud risk indicators, hidden transactional patterns, and anomalies that improved model performance.  Designed and deployed fraud detection and AML models using Graph Neural Networks (GNNs) to analyze complex transaction relationships and network behavior, significantly strengthening early risk detection and reducing fraud losses by 37%.  Enhanced enterprise credit risk assessment frameworks by improving ensemble-based machine learning models for Probability of Default (PD), Loss Given Default (LGD), and Exposure at Default (EAD), supporting more accurate and data-driven credit decisions.  Built time-series forecasting models and analytics dashboards to monitor market liquidity trends, detect emerging financial risks, and support scenario-based stress testing for proactive risk management.  Developed Retrieval-Augmented Generation (RAG) pipelines using GPT and LangChain to automate regulatory documentation workflows, improving knowledge retrieval and reducing Basel IV reporting preparation time by 55%.  Strengthened model governance and real-time fraud monitoring by integrating explainable AI tools (SHAP, LIME), MLOps frameworks (MLflow, Kubeflow), and scalable dashboards built with Streamlit and Plotly for operational risk analysis and decision support.

AI/ML Engineer

HCL Tech

2020-6 - 2022-7 · 2 yrs 2 mos

Project: AI-Driven COVID-19 Predictive Analytics and Resource Optimization Platform for Indian Healthcare Systems  Developed and maintained Python-based ETL pipelines using Pandas and NumPy to integrate and process large-scale patient and claims data from multiple Electronic Health Record (EHR) systems, enabling reliable healthcare analytics and reporting.  Standardized diverse healthcare datasets into the OMOP Common Data Model (CDM), improving data quality, interoperability, and enabling large-scale analytics across hospitals and public health programs.  Performed survival analysis and patient outcome modeling to identify risk patterns and support ICU capacity planning and critical resource allocation during peak COVID-19 surges.  Built and evaluated predictive machine learning models for patient readmission risk, improving prediction accuracy by 28% and supporting hospital planning for beds, ventilators, and clinical resources.  Implemented NLP pipelines using spaCy to extract symptoms, comorbidities, and clinical insights from unstructured medical notes while integrating ICD-10 and SNOMED standards for structured healthcare analysis.  Collaborated with clinicians and medical teams to interpret CNN-based medical imaging insights for COVID-19 pneumonia detection, ensuring model outputs were accurate, clinically relevant, and actionable.  Designed interactive Power BI dashboards and implemented HIPAA-compliant data governance frameworks, enabling physicians, epidemiologists, and health authorities to monitor disease trends and make data-driven decisions.