Generative AI and Ai/ml engineer

Vishnu Sai Reddy

Generative AI Engineer

+1-510-***-**** Email:***************@*****.*** LinkedIn: //www.linkedin.com/in/vishnu-bolla-b8ab411b2/

PROFESSIONAL SUMMARY

AI/ML Engineer with nearly 10 years of experience in Python development, Machine Learning, Deep Learning, data engineering, and MLOps across Azure, AWS, and GCP, with recent specialization in Generative AI, LLM applications, RAG, and Agentic AI.

Built Large Language Model (LLM) applications, knowledge assistants, and document workflows using Python, Azure OpenAI, GPT-4, GPT-4o, LangChain, LangGraph, and LlamaIndex.

Developed Retrieval-Augmented Generation (RAG) pipelines using Azure AI Search, Pinecone, and FAISS, covering ingestion, chunking, embeddings, hybrid retrieval, reranking, and grounded answers with source citations.

Built agentic workflows with tool calling state management, guardrails, and human review; explored Model Context Protocol (MCP)-style proofs of concept for controlled API and SQL access.

Delivered forecasting, recommendations, anomaly detection, and Natural Language Processing (NLP) solutions using SQL, PySpark, scikit-learn, XGBoost, TensorFlow, and PyTorch.

Built data preparation and feature-engineering pipelines using Azure Databricks, PySpark, Delta Lake, and Kafka to support AI applications and model development.

Evaluated LLM grounding and response quality, refining chunking, retrieval, reranking, and prompts while managing latency and token usage.

Worked across Azure, AWS, and GCP, applying MLOps and Azure-based LLMOps practices with FastAPI, MLflow, Docker, Kubernetes, and CI/CD for deployment, evaluation, monitoring, and production support.

Developed solutions for policy and compliance document search, energy forecasting, patient risk prediction, and retail customer analytics, connecting technical work to practical business needs.

Applied Responsible AI, access controls, audit logging, data privacy, and model explainability. Collaborated with business and engineering teams through delivery and mentored junior colleagues on feature engineering and model evaluation.

TECHNICAL SKILLS

Programming and Data: Python, SQL, PySpark, Pandas, NumPy, Apache Spark, Kafka, Delta Lake

Generative AI and Agents: Azure OpenAI, GPT-4, GPT-4o, LangChain, LangGraph, LlamaIndex, RAG, Prompt Engineering, Tool Calling, MCP-style POCs, LLM Evaluation

Retrieval and Search: Azure AI Search, Pinecone, FAISS, Embeddings, Hybrid Search, Reranking, Grounding

Machine Learning and NLP: Scikit-learn, XGBoost, TensorFlow, PyTorch, BERT, spaCy, Forecasting, Recommendations, Anomaly Detection

Cloud Platforms: Microsoft Azure: OpenAI, AI Search, Databricks, AKS; Amazon Web Services (AWS): SageMaker, S3, Glue, EKS; Google Cloud Platform (GCP): Vertex AI, BigQuery, Dataflow, Dataproc

APIs and Operations: FastAPI, Flask, REST APIs, Machine Learning Operations (MLOps), LLM Operations (LLMOps), MLflow, Docker, Kubernetes, Azure DevOps, GitLab CI/CD, Git

Databases and Reporting: SQL, BigQuery, Power BI, Tableau

AI Governance: Responsible AI, Access Controls, Grounding Checks, Audit Logging, Data Privacy, Model Explainability, Human-in-the-Loop Review

PROFESSIONAL EXPERIENCE

Mastercard Generative AI Engineer Arlington, VA March 2025 – Present

Project: Enterprise GenAI Knowledge & Document Intelligence Platform

Responsibilities:

As a Generative AI Engineer at Mastercard, I worked on an enterprise GenAI Knowledge and Document Intelligence Platform designed to help internal users securely search policies, retrieve relevant information, summarize documents, and automate knowledge-driven workflows.

Designed end-to-end RAG pipelines using Azure OpenAI and Azure AI Search, covering document ingestion, text extraction, semantic chunking, embeddings, hybrid retrieval, reranking, prompt grounding, and source citations.

Built LLM-powered knowledge assistants using GPT-4/4o and structured prompt templates for policy search, document Q&A, summarization, and enterprise knowledge retrieval.

Developed Agentic AI workflows using LangGraph, LangChain, and LlamaIndex for query routing, document retrieval, multi-document summarization, validation, and business-rule checks.

Implemented Pinecone and FAISS vector retrieval with metadata filtering and access-control attributes to support traceable and permission-aware document retrieval.

Created document intelligence workflows combining OCR, NLP, and LLMs for document extraction, classification, summarization, and question answering.

Developed MCP-style POCs for controlled integration with approved APIs and SQL tools, implementing schema validation, tool allowlisting, scoped permissions, audit logging, timeouts, and fallback handling.

Built FastAPI REST services to expose RAG, retrieval, and agent workflows and deployed containerized services on Azure Kubernetes Service (AKS) using Docker and Azure DevOps CI/CD.

Evaluated LLM grounding and response quality and optimized chunking, retrieval depth, reranking, and prompts to improve response relevance while managing latency and token usage.

Used MLflow for experiment and version tracking and maintained prompt versions, deployment validation, monitoring, and rollback procedures.

Implemented Responsible AI and security controls, including grounding checks, access controls, audit logging, model explainability, and human-in-the-loop review for sensitive workflows.

Collaborated with business, security, data, and QA teams on use-case definition, architecture reviews, release validation, and production support.

Environment: Python, SQL, Microsoft Azure, Azure OpenAI, GPT-4, GPT-4o, Azure AI Search, Azure Databricks, AKS, LangChain, LangGraph, LlamaIndex, Pinecone, FAISS, PySpark, Delta Lake, Kafka, FastAPI, REST APIs, MLflow, Docker, Kubernetes, Azure DevOps CI/CD, MCP-style POCs.

Champion Energy Gen Ai / Data Scientist Houston, TX September 2022 – February 2025

Project: AI-Powered Energy Analytics & Knowledge Intelligence Platform

Responsibilities:

As a Generative AI / Data Scientist at Champion Energy, I worked on an AI-powered energy analytics platform focused on demand forecasting, anomaly detection, customer analytics, and operational intelligence, with later-phase development of RAG-based knowledge assistance and Generative AI capabilities over approved process documents.

Built demand forecasting models using historical energy usage, weather signals, customer attributes, and operational data to support energy-demand planning and analysis.

Developed anomaly-detection models to identify unusual energy-usage and operational patterns and support exception analysis.

Built customer analytics and predictive models using Python, SQL, PySpark, Scikit-learn, and XGBoost to analyze customer behavior and energy-usage patterns.

Performed feature engineering, model validation, and performance evaluation to improve the reliability of forecasting and predictive models.

Prepared large-scale analytics datasets on GCP using BigQuery, Dataflow, Dataproc, and Apache Beam for machine-learning and analytics workloads.

Used Vertex AI to support model-development workflows and experimentation for energy forecasting and customer analytics use cases.

Automated energy-usage and customer-data ingestion and validation, creating curated datasets for recurring reporting, analytics, and ML development.

Developed a controlled RAG-based knowledge assistant during the later phase using LangChain, LlamaIndex, embeddings, and vector retrieval over approved process documents.

Evaluated Azure OpenAI and Llama models in later GenAI POCs for internal knowledge assistance, document retrieval, and summarization.

Designed and tuned document chunking and metadata-filtering strategies to improve retrieval quality across approved enterprise documentation.

Optimized query expansion, top-k retrieval, retrieval depth, and prompt strategies and validated knowledge-assistant changes before release to business users.

Applied BERT and spaCy for text classification, entity extraction, and analysis of customer communications and other unstructured business data.

Used MLflow for experiment tracking and model evaluation and GitLab CI/CD to support controlled and repeatable application and model releases.

Developed Tableau and Power BI dashboards to communicate forecasting results, customer analytics, model outputs, operational KPIs, and data-quality information.

Collaborated with operations, product, and data teams to define requirements, validate model results, evaluate AI use cases, support business releases, and mentor junior colleagues on feature engineering and model evaluation.

Environment: Python, SQL, GCP, Vertex AI, BigQuery, Dataflow, Dataproc, Apache Beam, PySpark, scikit-learn, XGBoost, BERT, spaCy, LangChain, LlamaIndex, MLflow, GitLab CI/CD, Tableau, Power BI, Azure OpenAI and Llama model POCs in the later phase.

NantHealth AI/ML Engineer / Data Scientist Winterville, NC April 2020 – August 2022

Project: AI-Powered Clinical Analytics & Patient Risk Prediction Platform

Responsibilities:

As an AI/ML Engineer / Data Scientist at NantHealth, I worked on a healthcare analytics platform focused on patient risk prediction, readmission analysis, clinical NLP, medical coding support, and healthcare data intelligence, using machine learning, deep learning, and NLP technologies on AWS.

Developed patient risk prediction and hospital readmission models using Scikit-learn, XGBoost, TensorFlow, and PyTorch to support healthcare analytics and clinical decision-support workflows.

Performed feature engineering, cross-validation, hyperparameter tuning, and model evaluation to improve the reliability and generalization of predictive healthcare models.

Built clinical NLP pipelines using BERT and spaCy for clinical text classification, entity extraction, medical coding support, and information retrieval from EHR notes and clinical documents.

Prepared structured and unstructured healthcare datasets using Python, SQL, PySpark, Amazon S3, and AWS Glue for model training, analytics, and downstream applications.

Developed data preprocessing and validation workflows to handle clinical and patient-related data, improving the consistency of datasets used for model development.

Supported model training and deployment workflows using Amazon SageMaker, enabling repeatable development and operationalization of machine-learning models.

Containerized ML applications using Docker and supported scalable deployment on Amazon EKS/Kubernetes with CI/CD-based release workflows.

Used MLflow for experiment tracking, model comparison, and model lifecycle management across healthcare machine-learning use cases.

Developed CNN-based image classification models for selected healthcare imaging and document-image use cases and validated model results before application integration.

Exposed ML model predictions through FastAPI and Flask REST services, enabling downstream healthcare applications to consume predictive outputs.

Applied model explainability, privacy safeguards, access controls, and HIPAA-aligned data handling to support responsible use of healthcare AI/ML solutions.

Collaborated with analysts, data engineers, and application teams on model testing, inference monitoring, production support, retraining workflows, and Power BI reporting of healthcare and model-performance KPIs.

Environment: Python, SQL, AWS, Amazon SageMaker, Amazon S3, AWS Glue, Amazon EKS, PySpark, Scikit-learn, XGBoost, TensorFlow, PyTorch, BERT, spaCy, FastAPI, Flask, MLflow, Docker, Kubernetes, CI/CD, Power BI.

Lululemon Data Scientist / ML Engineer Dallas, TX January 2018 – March 2020

Project: Customer Intelligence & Retail Recommendation Analytics Platform

Responsibilities:

As a Machine Learning / Data Scientist at Lululemon, I worked on a retail customer intelligence platform focused on customer behavior analytics, churn prediction, segmentation, personalized recommendations, demand forecasting, and inventory analytics using machine learning and statistical modeling.

Developed customer churn prediction models using Python, Scikit-learn, and XGBoost to identify customers with higher likelihood of disengagement based on shopping and engagement behavior.

Built customer segmentation models using behavioral, purchase, and engagement data to identify meaningful customer groups for retail analytics and personalization.

Compared Random Forest, Logistic Regression, and XGBoost models using customer behavior features to evaluate and improve churn-prediction performance.

Developed recommendation models using collaborative filtering, product attributes, and purchase history to support personalized product recommendations.

Built demand forecasting models using historical sales, seasonality, product, and store-level information to support retail planning and inventory analysis.

Processed large-scale retail datasets using SQL, PySpark, Apache Spark, and Databricks, preparing repeatable datasets for machine-learning and analytics workloads.

Performed data cleaning, feature engineering, exploratory data analysis, and statistical analysis to identify customer, product, and purchasing patterns.

Engineered customer-behavior and product-level features to support churn, recommendation, segmentation, and demand-forecasting models.

Applied spaCy, NLTK, sentiment analysis, and text classification to customer feedback and other unstructured retail text.

Developed anomaly-detection analysis to identify unusual sales and inventory patterns and support investigation of retail exceptions.

Evaluated machine-learning models through cross-validation and business review and used MLflow for selected experiment-tracking workflows during the later phase of the engagement.

Developed Tableau and Power BI dashboards and lightweight Flask scoring services, collaborating with merchandising, BI, and data teams to make customer and model insights usable by business stakeholders.

Environment: Python, SQL, scikit-learn, XGBoost, Random Forest, Logistic Regression, PySpark, Apache Spark, Databricks, NLTK, spaCy, Flask, Tableau, Power BI, MLflow for selected workflows in the later phase.

EPAM Systems Python Developer Hyderabad, India July 2016 – June 2017

Project: Enterprise Data Processing & Reporting Automation Platform

Responsibilities:

As a Python Developer at EPAM Systems, I worked on enterprise data-processing and reporting solutions focused on Python-based data extraction, cleaning, transformation, validation, SQL reporting, and large-scale data processing, supporting business analytics and operational reporting requirements.

Developed reusable Python and SQL workflows for data extraction, cleaning, profiling, transformation, and validation using Pandas and NumPy.

Built and maintained SQL queries and reporting datasets to support recurring business analytics and reporting requirements.

Developed reusable reporting datasets for finance, procurement, HR, and operational analytics, validating data and results with business stakeholders.

Supported large-scale data processing using Apache Spark, PySpark, Hadoop, and Hive for enterprise datasets.

Assisted with predictive modeling and exploratory data analysis using Scikit-learn and XGBoost, including model evaluation and data preparation.

Developed classical NLP workflows for keyword extraction, document categorization, text classification, and information extraction from business documents.

Created Power BI and Tableau reports, documented Python and data-processing logic, and maintained source code through Git and Linux-based development workflows.

Investigated data-quality issues and failed processing jobs and collaborated with senior developers, data engineers, and business users to troubleshoot issues and provide production support.

Environment: Python, SQL, Pandas, NumPy, scikit-learn, XGBoost, Apache Spark, PySpark, Hadoop, Hive, Power BI, Tableau, Git, Linux.

EDUCATION

Bachelor’s in computer science Amity University 2016

Contact this candidate