Home

Guna Raj Nedunuru - Senioe Genrative Engineer

[email protected]

Location: San Jose, California, USA

Relocation:

Visa: H1B

Guna Raj Nedunuru Sr. Generative AI Engineer [email protected] | +1 7193627749 | Linkedin PROFESSIONAL SUMMARY: Senior Generative AI Engineer with 10 years of experience building enterprise AI and cloud-native platforms focused on Generative AI, LLMs, RAG, NLP, and intelligent search solutions across AWS and Azure ecosystems. Experienced in developing production-grade AI applications that improve knowledge retrieval, automate business workflows, and support scalable enterprise adoption of AI technologies. Expertise in developing and optimizing LLM-powered applications such as conversational AI, enterprise virtual assistants, intelligent document summarization, question-answering systems, and AI-driven search platforms with a focus on contextual relevance, response accuracy, and low-latency inference. Experienced in implementing LLMOps and MLOps best practices, including model versioning, prompt lifecycle management, CI/CD automation, experiment tracking, monitoring, evaluation pipelines, automated retraining, and multi-environment deployment strategies for stable production AI systems. Proficient in Python, SQL, TensorFlow, PyTorch, FastAPI, Flask, MLflow, Docker, and Kubernetes with strong experience building scalable machine learning pipelines, RESTful APIs, microservices, and cloud-native AI applications. Skilled in Responsible AI and enterprise AI governance with experience implementing hallucination mitigation, PII masking, content moderation, RBAC, encryption, secure APIs, and compliance-focused AI deployment standards across enterprise environments. TECHNICAL SKILLS: Programming & Frameworks: Python, SQL, Pandas, NumPy, SciPy, PySpark, TensorFlow, PyTorch, Flask, FastAPI, Matplotlib, Seaborn Generative AI / NLP: Generative AI, LLMs, RAG, Prompt Engineering, Semantic Search, Conversational AI, Vector Search, NLP, Transformer Models, Embeddings, LLMOps, Intelligent Search, AI Governance Machine Learning / AI: Machine Learning, Deep Learning, Neural Networks, Predictive Analytics, Recommendation Systems, Classification, Regression, Forecasting, Supervised & Unsupervised Learning, Feature Engineering, Hyperparameter Optimization, Model Evaluation, Model Monitoring, Automated Retraining, Real-Time & Batch Inference, Time-Series Forecasting, A/B Testing Cloud & AI Services: AWS, Amazon Bedrock, SageMaker, AWS Lambda, Step Functions, S3, ECS, EKS, ECR, IAM, KMS, CloudWatch, OpenSearch, Azure Data Factory, Azure Data Lake, Azure Blob Storage, AKS, Azure DevOps, Azure Monitor, Application Insights, Azure Active Directory, Azure Key Vault, Azure Machine Learning DevOps / MLOps: MLOps, CI/CD, Docker, Kubernetes, MLflow, Model Versioning, Experiment Tracking, Automated Testing, Release Management, Monitoring & Alerting, Autoscaling, Logging, Unit Testing APIs / Data Engineering: REST APIs, Microservices, ETL Pipelines, Data Ingestion, Data Processing, Data Transformation, Feature Stores, Batch Processing, Event-Driven Architecture, Serverless Architecture Security / Governance: RBAC, Secure APIs, Encryption at Rest & In Transit, IAM Policies, Secrets Management, Compliance, Risk Management, Authentication & Authorization Methodologies: Agile, Scrum, Sprint Planning, Backlog Refinement, Architecture Reviews, Release Coordination, Technical Documentation, Operational Runbooks, Cross-Functional Collaboration PROFESSIONAL EXPERIENCE: Comerica Bank, Dallas TX | November 2024 Present Senior Generative AI Engineer Responsibilities: Architected and deployed a ComerIQ-aligned GenAI platform for the financial-services domain, enabling ingestion, embedding generation, vector indexing, prompt orchestration, and low-latency inference across 12M+ enterprise documents. Designed production RAG architectures using proprietary banking content and regulatory knowledge sources to power an internal knowledge copilot, improving answer grounding for compliance, operations, and technology teams. Built reusable prompt-engineering workflows for IT and colleague-facing copilots, including template libraries, automated evaluation, and feedback-driven tuning that improved response consistency and reduced hallucination risk. Led GenAI delivery for Commercial Banking, Treasury, and Technology teams, automating knowledge retrieval, document analysis, intelligent search, and decision-support workflows to shorten manual research cycles. Defined enterprise LLMOps standards for prompt lifecycle management, model versioning, evaluation gates, rollback procedures, release governance, and multi-environment deployment of banking AI services. Implemented Responsible AI guardrails for financial-services use cases, including input validation, PII masking, toxicity filtering, grounded-response controls, and output moderation aligned with regulatory expectations. Integrated GenAI capabilities with internal banking platforms through secure REST APIs and role-based access controls, enabling governed adoption across Commercial Banking, Retail, IT, and Operations. Led architecture reviews and mentored engineering teams on scalable GenAI patterns, deployment practices, and AI governance controls aligned with the bank's enterprise technology strategy. Delivered LLM-powered products for the banking domain, including policy semantic search, a colleague self-service virtual assistant, and document summarization tools for Treasury and Compliance workflows. Implemented high-performance vector retrieval with Amazon OpenSearch and embedding-based search to provide low-latency access to large banking policy and operational document repositories. Performed inference benchmarking, token-usage analysis, and capacity optimization to improve throughput, control infrastructure cost, and maintain production service-level objectives. Operationalized Amazon Bedrock foundation models for text generation, conversational AI, embedding generation, and enterprise intelligent-search capabilities. Built scalable S3-based ingestion and embedding pipelines to manage enterprise datasets, vector-index inputs, and model artifacts for RAG workloads. Developed AWS Lambda functions for document preprocessing, prompt orchestration, and response post-processing, reducing operational overhead for event-driven AI inference services. Orchestrated end-to-end GenAI workflows with AWS Step Functions for ingestion, embedding creation, retrieval, inference, and governed response delivery. Containerized and deployed FastAPI-based GenAI microservices on Amazon ECS and EKS, supporting scalable and highly available inference for enterprise banking applications. Established CloudWatch observability for latency, token utilization, error rates, and infrastructure health, reducing incident response time by 35% and improving platform stability. Applied AWS IAM least-privilege policies and KMS encryption to protect sensitive banking data in RAG and AI services, supporting enterprise compliance and governance requirements. Partnered with legal, compliance, privacy, and cybersecurity stakeholders to translate banking governance requirements into deployable GenAI controls and operating standards. Environment: Amazon Bedrock, AWS Lambda, AWS Step Functions, Amazon ECS, Amazon EKS, Amazon S3, AWS OpenSearch, Amazon CloudWatch, AWS IAM, AWS KMS,Python, FastAPI, LangChain, LlamaIndex, RAG, Vector Embeddings, Prompt Engineering, LLMOps, Docker, Kubernetes, CI/CD, MLflow, Git, Responsible AI: PII Masking, Hallucination Mitigation, RBAC, Content Moderation Health First, Rockledge, Florida | March 2023 October 2024 AI/ML Engineer Responsibilities: Built scalable ML data pipelines for the healthcare domain, processing structured and unstructured EHR, clinical-note, and patient-risk data through standardized preprocessing, transformation, and feature-engineering workflows. Developed TensorFlow and PyTorch models for clinical NLP, readmission-risk prediction, appointment no-show forecasting, and intelligent billing classification in support of population-health operations. Delivered predictive models for inpatient trajectory risk scoring and high-risk outpatient identification, supporting clinical and care-management workflows across Health First's four-hospital health system. Designed model evaluation, monitoring, and automated retraining frameworks to detect data drift, sustain model reliability, and improve clinical decision-support performance across health-plan operations. Integrated ML services with enterprise ETL and EHR data platforms to enable reliable, scalable, and low-latency clinical data processing. Standardized MLOps practices with CI/CD automation, model versioning, experiment tracking, and automated validation, improving repeatability and release confidence for healthcare AI applications. Implemented event-driven AWS Lambda preprocessing and inference services for low-latency clinical predictions, billing workflows, and appointment-scheduling operations. Optimized model serving through right-sized compute, autoscaling, and efficient deployment patterns, improving inference efficiency and reducing cloud operating costs. Trained, tuned, deployed, and managed production models on Amazon SageMaker using managed training jobs, hyperparameter optimization, and scalable real-time endpoints. Built clinical NLP pipelines for medical-text preprocessing, entity extraction, tokenization, embedding generation, and transformer inference over unstructured EHR content. Fine-tuned transformer and foundation models for healthcare document summarization, clinical Q&A, and ICD code suggestions, improving domain relevance across clinician workflows. Configured CloudWatch logging, monitoring, and alerting to track model quality, system reliability, and application health across the AI platform. Collaborated with clinicians, health-plan leaders, and IT teams to convert clinical requirements into measurable AI solutions that improved operational efficiency and supported patient-care decisions. Orchestrated AWS Step Functions workflows for automated training, batch inference, and model updates supporting patient-data and healthcare supply-chain processes Environment: Amazon SageMaker, AWS Lambda, AWS Step Functions, Amazon ECR, Amazon ECS, Amazon EKS, Amazon S3, Amazon CloudWatch, AWS IAM, AWS KMS, Python, TensorFlow, PyTorch, Scikit-learn, Hugging Face Transformers, NLP, EHR Data Processing, Docker. Onward Technologies, Hyderabad, India | June 2019 November 2022 Machine Learning Engineer Responsibilities: Designed Azure Data Factory orchestration and managed enterprise datasets in ADLS Gen2 and Blob Storage for industrial, automotive, and healthcare engineering clients. Containerized ML applications with Docker and deployed highly available microservices on AKS, enabling scalable production inference for digital-engineering workloads. Built Azure DevOps CI/CD pipelines to automate model integration, testing, deployment, and release management across Industry 4.0 and Digital Factory programs. Engineered predictive-maintenance models for industrial equipment and automotive OEMs using sensor and telemetry data to forecast component failures and optimize maintenance schedules. Developed end-to-end ML pipelines for batch and near-real-time inference across transportation, mobility, industrial equipment, and healthcare MedTech client programs. Created Flask and FastAPI REST APIs to expose ML inference services and embed predictive capabilities into downstream automotive and industrial enterprise applications. Managed model lifecycle operations in Azure Machine Learning, including experiment tracking, model registry, deployment approvals, and managed online endpoints. Built reusable feature-engineering and data-preprocessing frameworks with Python, Pandas, NumPy, and PySpark for high-volume manufacturing and engineering datasets. Implemented MLflow for experiment tracking, model lifecycle management, artifact storage, and version control across digital-engineering analytics workflows. Strengthened multi-client cloud security with Azure Active Directory authentication, RBAC, and Azure Key Vault-based secrets management. Partnered with data scientists and OEM domain experts to productionize research models for automotive ADAS, embedded systems, industrial engineering, and healthcare MedTech solutions. Contributed to Agile delivery through sprint planning, backlog refinement, code reviews, and quality practices within TISAX-certified and ISO 9001-compliant engineering programs. Environment: Azure Data Factory, Azure Data Lake Storage Gen2, Azure Blob Storage, Azure Kubernetes Service (AKS), Azure Machine Learning, Azure DevOps, Azure Monitor, Application Insights, Azure Active Directory, Azure Key Vault, Python, PySpark,Docker, Kubernetes. KPIT Technologies, India | January 2015 May 2019 Data Scientist Responsibilities: Performed data preparation, feature engineering, and statistical analysis with Python, Pandas, NumPy, and SciPy to build predictive models for automotive OEM vehicle diagnostics, telematics, and software-defined vehicle programs. Extracted, transformed, and validated high-volume automotive telemetry and operational data using complex SQL, enabling vehicle-performance analytics and powertrain reporting. Conducted exploratory data analysis with Matplotlib and Seaborn to identify trends, anomalies, and behavioral patterns in automotive sensor data for OEM forecasting and validation. Designed Remaining Useful Life (RUL) predictive-maintenance models for vehicle components, helping automotive OEM and Tier-1 engineering teams anticipate failure risk and optimize service planning. Built scenario-driven drive-log analytics pipelines that detected critical events for ADAS validation, testing, and verification coverage. Developed and optimized supervised and unsupervised models using Linear and Logistic Regression, Decision Trees, Random Forest, Gradient Boosting, K-Means, and Naive Bayes for vehicle diagnostics and operational analytics. Automated recurring engineering reports and analytical workflows with Python scripts and scheduling, reducing manual processing for OEM stakeholder teams. Created ARIMA, moving-average, and trend-analysis forecasting models for vehicle demand, capacity planning, and component lifecycle management across automotive production programs. Evaluated model performance with cross-validation, ROC-AUC, accuracy, precision, recall, F1-score, and confusion matrices to ensure reliable analytics outcomes. Collaborated with automotive engineers, OEM analysts, and business stakeholders to translate vehicle-software requirements into scalable analytics and predictive-modeling solutions. Applied hypothesis testing and A/B testing to validate model recommendations, measure effectiveness, and quantify business impact for data-driven automotive engineering initiatives. Environment: Python, Pandas, NumPy, SciPy, Matplotlib, Seaborn, Scikit-learn, SQL, Relational Databases, ETL Pipelines, Machine Learning: Random Forest, Gradient Boosting, K-Means, ARIMA, Time-Series Forecasting, Industry Domain: Automotive Software, ADAS Validation, Predictive Maintenance, Vehicle Diagnostics, Software-Defined Vehicles (SDV) EDUCATION: Bachelors in Computer Science National Institute of Technology Warangal, India Keywords: continuous integration continuous deployment artificial intelligence machine learning sthree information technology Texas