Home

Abhinash Reddy - AI ML Engineer

[email protected]

Location: Remote, Remote, USA

Relocation: YES

Visa: GC

Resume file: Abhinash_AI_ML_Engineer_1764784729196.docx Please check the file(s) for viruses. Files are checked manually and then made available for download.

Abhinash Reddy Peddyreddy Sr.AI/ML Engineer [email protected] (270) 226-0303 Professional Summary AI/ML Engineer with 11+ years of experience delivering enterprise-grade AI, ML, and Generative AI solutions using Python, R, and Java across Cloud, Healthcare, Telecom, and FinTech domains. Strong expertise in Large Language Models (LLMs) including GPT-based and multilingual prompt engineering, applied in production-grade NLP solutions. Demonstrated success in building and benchmarking GenAI systems across platforms like AWS Bedrock, Azure OpenAI, and Vertex AI. Developed predictive models using GLM, Random Forest, XGBoost, and neural networks for financial risk detection, anomaly detection, and forecast optimization. Extensive hands-on experience with OpenAI, Google Gemini, Azure OpenAI, Azure ML Studio, and Prompt Flow for building cutting-edge AI pipelines. Proven expertise in crafting production-grade prompts, deploying RAG frameworks, and integrating GenAI capabilities for real-time customer-facing applications. Skilled in building ML pipelines for applications like OCR (handwritten), anomaly detection, knowledge-based chatbots, and adaptive learning systems. Proficient in natural language processing (NLP), text mining, information extraction, language generation, and decision science models. Advanced development with Apache Spark ecosystem including Spark Core, Spark SQL, Spark Streaming, and MLlib. Expert in regression analysis (linear, logistic, Poisson, binomial), deep learning (ANN, DNN), and probabilistic models. Strong programming expertise in Python with libraries like NumPy, Scikit-learn, Gensim, NLTK, TensorFlow, Keras. Skilled in ETL pipeline development, data modeling, data architecture, and end-to-end machine learning system productionization. Proven ability in building analytics POCs, working with big data (Spark, Scala, PySpark), and implementing data-driven solutions. Deep understanding of globalization/localization, language readiness feedback, and translation validation, with hands-on experience in software internationalization. Experience in SQL, data mining algorithms, PowerShell scripting, and working with various file systems, servers, and databases. Certified and experienced in cloud platforms including Azure, with knowledge of CI/CD pipelines and scalable ML deployment best practices. Technical Skills: Programming Languages Python, JAVA, JavaScript, jQuery, ReactJS, Next.js, HTML, CSS, C, C++, Angular, R, Impala, Hive, SQL Statistical Methods Statistical Inference, Hypothesis Testing, Confidence Intervals, p-values, Statistical Significance, Probability Distributions, Descriptive Statistics, Correlation and Covariance, Sampling Techniques, ANOVA, Chi-Square Tests, Bayes' Theorem, Cross-Validation, Time Series Analysis, Auto-correlation, Statistical Modelling, Experimental Design, Central Limit Theorem, Law of Large Numbers, Residual Analysis, Multivariate Analysis Machine Learning Supervised Learning, Unsupervised Learning, Model Evaluation Metrics, Cross-Validation, Feature Engineering, Hyperparameter Tuning, Overfitting & Regularization (L1, L2), Ensemble Methods, Decision Trees, Random Forest, Gradient Boosting (XGBoost, LightGBM), Support Vector Machines (SVM), K-Nearest Neighbors (KNN), Naive Bayes, Regression (Linear, Logistic), Classification Algorithms, Clustering (K-Means, Hierarchical), Dimensionality Reduction (PCA, t-SNE), Time Series Forecasting, ARIMA, SARIMA, Holt-Winters Exponential Smoothing, Prophet, LSTM for Time Series, Model Interpretability (SHAP, LIME), Recommendation Systems, Scikit-learn, Azure ML Studio, Azure Machine Learning Services, LLM Fine-tuning, Multi-Agent Systems, Retrieval-Augmented Generation (RAG) Deep Learning Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Autoencoders, Transfer Learning(VGG, ResNet, InceptionNet, MobileNet), Attention Mechanism, Object Detection (YOLO, Faster R-CNN), OCR, Image Segmentation (U-Net, Mask R-CNN), Activation Functions (ReLU, Sigmoid, Tanh, Softmax), Optimization Algorithms (SGD, Adam, RMSprop), Loss Functions (Cross-Entropy, MSE), Regularization (Dropout, Batch Normalization), Model Deployment, TensorFlow, Keras, PyTorch Natural Language Processing (NLP) Text Preprocessing (Tokenization, Lemmatization, Stemming), Bag of Words (BoW), TF-IDF, Word Embeddings (Word2Vec, GloVe, FastText), Sequence-to-Sequence Models, Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), Attention Mechanism, Encoder-Decoder Models, Gated Recurrent Units (GRU), Transformer Models (BERT, GPT, T5), Named Entity Recognition (NER), Part-of-Speech Tagging, Sentiment Analysis, Text Classification, Machine Translation, Text Summarization, Question Answering Systems, Language Modeling, Word and Sentence Similarity, POS Tagging, Dependency Parsing, Text Generation, Sequence Labeling, Topic Modeling (LDA), SpaCy, NLTK, Hugging Face Generative AI LLM s, Ollama, Langchain, Langsmith, Agentic AI, Fine-tuning Techniques (LoRA, QLoRA), Local LLMs (Mistral, Llama, Gemma, etc.), Inference APIs (Groq, Nvidia NIM, etc.), Retrieval-Augmented Generation (RAG), Graph Databases(Neo4j), Vector Databases (Faiss, Chroma, Pinecone), Embedding Models(Mistral, GoogleAI Embedding Models), Prompt Engineering, Prompt Templates, Prompt Versioning, Document Chunking & Indexing, Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), Conditional GANs, CycleGAN, StyleGAN, Deep Convolutional GANs (DCGAN), Transformer-based Generative Models (GPT, T5), Text-to-Image Generation (DALL-E, CLIP), Image-to-Image Translation, Neural Style Transfer, Chatbots, AI Search Engines, Semi-Supervised Learning, Data Augmentation, Zero-Shot Learning, Self-Supervised Learning, Reinforcement Learning for Generation, Diffusion Models, Latent Space Exploration, LangChain, Vector Databases, Embedding Models MLOps & LLMOps CI/CD for ML, Model Deployment (REST API, Batch), Model Monitoring, Model Versioning, Model Retraining Pipelines, ML Pipelines (MLflow, Airflow, Kubeflow), Experiment Tracking (MLflow, Weights & Biases), Data Versioning (DVC), Logging and Alerting (Prometheus, Grafana), Containerization (Docker), Orchestration (Kubernetes), Feature Stores (Feast), Cloud ML Platforms (AWS SageMaker, AWS BedRock, GCP AI Platform, Google Vertex AI, Azure AI) LLM Response Evaluation (TruLens, PromptLayer), Model Registry, Token Usage Tracking, Scalable LLM Deployment (TGI, vLLM), Azure DevOps, Model Deployment on Azure, Model Versioning, Monitoring & Retraining Pipelines Big Data: Hadoop, Hive, HBase, Apache Spark, Scala, Kinesis, Pig, Sqoop, Databricks, Apache Spark, Azure Data Lake, Azure Synapse, Azure Blob Storage Amazon Web Services: EC2, Lambda, Sage Maker, Bedrock, EMR, S3, Glue, MKS, Kinesis, Quick Sight, API Gateway, Athena, Lex, Recognition, CI/CD, Code Commit, DynamoDB, transcribe, Cloud Formation, Cloud Watch, Glacier, IAM Database Servers: MySQL, Microsoft SQL server, SQLite, Red Shift, RDS, PostgreSQL, SQL Alchemy, Mongo DB, Teradata Other Tools & Technologies: Git, GitHub, GitLab, Docker, Docker Compose, Kubernetes, VS Code, Jupyter Notebook, Google Colab, CUDA Toolkit, Postman, Swagger, REST APIs, Linux/Unix Command Line, Bash Scripting, Conda, Virtualenv, Makefile, YAML, JSON, Terraform, Nginx, Redis, Jenkins, Power BI, Tableau, Excel, Google Sheets, Notion, Trello, Slack, Figma Professional Experience: Client: Oracle, Austin, Texas Dec 2023 Present Role: Sr. AI/ML Engineer Project: Oracle AI Cloud Services Intelligent Automation and Generative AI Platform Responsibilities: Led development of Generative AI solutions within Oracle Cloud Infrastructure (OCI), integrating LLMs for intelligent automation, customer engagement, and system diagnostics across enterprise applications. Designed and implemented Retrieval-Augmented Generation (RAG) frameworks leveraging Oracle AI Vector Search, LangChain, and FAISS to enhance document retrieval and enterprise search within Oracle Fusion Cloud modules. Built multi-agent GenAI systems using LangChain Agents and Vertex AI to automate knowledge base curation, service ticket classification, and internal IT workflow support. Evaluated and fine-tuned large language models (GPT-4, Claude, Gemini, Llama-2) on Oracle Cloud GPU instances to optimize latency, prompt fidelity, and domain accuracy for cloud operations use cases. Developed NLP pipelines using BERT, spaCy, and Hugging Face Transformers for intelligent document processing, entity recognition, and knowledge extraction from Oracle support tickets and policy docs. Engineered AI microservices using FastAPI and Flask, exposing RESTful APIs for anomaly detection, forecasting, and chat-based support assistants integrated with Oracle Digital Assistant platforms. Implemented MLOps and LLMOps pipelines with MLflow, Airflow, and OCI Data Science Services to streamline model training, deployment, and continuous governance across multi-cloud environments. Built and deployed predictive models (XGBoost, LightGBM, PyTorch) for capacity planning, system health prediction, and incident response optimization, reducing mean time to resolution by 25%. Designed real-time data ingestion and monitoring frameworks using Kafka, Spark Streaming, and Delta Lake, processing billions of log events and telemetry records daily. Constructed feature stores and data pipelines in PySpark and Snowflake, integrated with OCI Data Integration to enable scalable model training on enterprise datasets. Created dashboard visualizations in Grafana and Oracle Analytics Cloud (OAC) to track model KPIs, inference latency, and AI adoption metrics across cloud services. Established LLM evaluation and monitoring frameworks using TruLens and PromptLayer, ensuring explainability and responsible AI alignment with Oracle s trust framework. Collaborated with cloud platform and security teams to integrate AI systems with OCI Identity and Access Management (IAM) and Vault services for secure key management and data protection. Developed IaC templates with Terraform and Ansible to automate deployment of AI infrastructure across staging and production environments. Partnered with Oracle Fusion and ERP product teams to embed GenAI-based assistants within cloud applications for report summarization, query generation, and recommendation features. Conducted A/B tests and LLM performance audits to measure prompt accuracy and user satisfaction, improving response precision by 20%. Mentored data science and DevOps teams on modern AI tooling, GenAI integration, and model lifecycle automation within Oracle s AI platform ecosystem. Played a key role in establishing Oracle s enterprise-grade GenAI Center of Excellence, defining governance standards and deployment templates for AI-enabled cloud services. Designed and deployed RAG pipelines and multi-agent LLM systems using Azure ML Studio and Azure Cognitive Search for enterprise knowledge management. Integrated AI/ML models with Azure Data Lake, Azure Synapse, and Azure Blob Storage for seamless data ingestion and processing Implemented MLOps pipelines using Azure DevOps for model versioning, deployment, monitoring, and automated retraining Applied LangChain and vector database embeddings for retrieval and semantic search in large-scale enterprise datasets. Environment: Python, PySpark, Pandas, NumPy, Scikit-learn, XGBoost, LightGBM, TensorFlow 2.x, PyTorch, BERT, Hugging Face Transformers, LangChain, Oracle Cloud Infrastructure (OCI Data Science, AI Services, Vector Search), Vertex AI, FastAPI, Flask, MLflow, Airflow, Docker, Kubernetes (OKE), Terraform, Ansible, Snowflake, Delta Lake, Kafka, Spark Streaming, Grafana, Oracle Analytics Cloud (OAC), Git, Jenkins, JSON, YAML, Linux, Agile/Scrum. Client: Paypal, CA(Remote) Apr 2021 Dec 2023 Role: Sr. AI/ML Engineer Project: Generative AI Driven Fraud Intelligence and Customer Experience Automation Platform Responsibilities: Led design and deployment of Generative AI systems for fraud analytics, transaction monitoring, and conversational support, leveraging LangChain, Vertex AI, and AWS Bedrock. Architected Retrieval-Augmented Generation (RAG) pipelines integrating PayPal s internal data lake with vector databases (Pinecone, FAISS) to support semantic search and fraud case summarization. Built and productionized multi-agent GenAI systems using Google Agent Builder and LangChain Agents to automate compliance documentation review, KYC/AML validations, and transaction alerts. Conducted LLM benchmarking (GPT-4, Claude, Gemini, Llama-2) to assess precision, latency, and retrieval quality across financial datasets. Integrated AWS Comprehend, Bedrock, and Vertex AI Search to extract insights from unstructured data sources like customer messages, chargeback forms, and support transcripts. Developed NLP pipelines for intent detection, sentiment analysis, and fraud claim classification using spaCy, BERT, and Hugging Face Transformers. Deployed RAG-powered virtual assistants and chatbots for customer dispute resolution and fraud escalation workflows, reducing manual triage time by 38%. Built ML pipelines for anomaly and fraud detection using XGBoost, LightGBM, and PyTorch, enhancing detection accuracy by 26% with real-time inference capabilities. Designed streaming data ingestion frameworks using Kafka, Spark Streaming, and Delta Lake to process millions of payment events daily with sub-second latency. Developed RESTful AI microservices using FastAPI and Flask, exposing model APIs for transaction scoring, chatbot responses, and fraud predictions. Created MLOps and LLMOps pipelines using MLflow, Airflow, SageMaker, and Vertex AI Pipelines for automated model versioning, evaluation, and CI/CD deployment. Integrated TruLens, PromptLayer, and RLHF frameworks to evaluate LLM outputs, monitor bias, and fine-tune prompt templates based on PayPal compliance standards. Collaborated with risk and data governance teams to ensure PCI-DSS, GDPR, and AML/KYC compliance in all AI workflows and GenAI systems. Optimized data ingestion, feature generation, and storage using Snowflake, Delta Lake, and PySpark, reducing compute cost and data latency by 30%. Engineered infrastructure-as-code (IaC) using Terraform and AWS CloudFormation for scalable deployment of GenAI and ML workloads. Partnered with fraud investigation units to create graph-based link analysis models using Neo4j and GraphSAGE, identifying fraudulent transaction networks in real time. Built end-to-end dashboards in Tableau and Grafana to visualize model KPIs, risk thresholds, and GenAI output metrics for compliance audits. Automated LLM retraining triggers via Airflow DAGs based on data drift detection and user feedback loops. Conducted A/B testing and impact analysis to validate model effectiveness in reducing fraud rates, false positives, and customer churn. Mentored junior AI engineers and data scientists in MLOps best practices, LLM fine-tuning, and agent orchestration frameworks. Played a key role in PayPal s transition toward AI-driven risk intelligence and GenAI-enabled customer automation, aligning innovation with security and regulatory goals. Extended GenAI fraud and compliance systems to Azure ML environments, leveraging multi-agent orchestration, vector databases, and embedding models. Built end-to-end ML pipelines on Databricks and Apache Spark, integrating Azure-based data services for high-performance model training and inference. Environment: Python, PySpark, Pandas, NumPy, Scikit-learn, XGBoost, LightGBM, TensorFlow 2.x, PyTorch, BERT, Hugging Face Transformers, LangChain, Vertex AI, AWS SageMaker, AWS Bedrock, AWS Comprehend, Google Agent Builder, FastAPI, Flask, MLflow, Airflow, Docker, Kubernetes (EKS), Terraform, Snowflake, Delta Lake, Spark Streaming, Kafka, Pinecone, FAISS, Neo4j, Grafana, Tableau, Git, GitHub Actions, Jenkins, JSON, YAML, Linux, Agile/Scrum. Client: Verizon Communications, Irving, Texas Nov 2019 Mar 2021 Role: AI/ML Engineer Project: AI-Driven Network Optimization & Customer Retention Platform Responsibilities: Spearheaded the design and deployment of AI models for network fault prediction, churn analysis, and service outage detection, enabling proactive maintenance and customer retention. Built end-to-end machine learning pipelines in AWS SageMaker and GCP Vertex AI, orchestrating model training, validation, and deployment across multiple network monitoring systems. Developed predictive models using XGBoost, LSTM, and Prophet to forecast network traffic, service degradation, and subscriber churn probability with 92% accuracy. Implemented real-time anomaly detection systems leveraging Kafka streams, Spark Structured Streaming, and PyTorch, reducing false alarms by 28%. Integrated telemetry and IoT sensor data from towers and routers into centralized data lakes for time-series forecasting and fault diagnostics. Designed NLP-based analytics pipelines using BERT and spaCy to analyze customer feedback, call transcripts, and complaint logs for sentiment and intent classification. Developed RESTful inference APIs using FastAPI for real-time predictions embedded in network operations dashboards. Containerized ML models using Docker and deployed on Kubernetes (Amazon EKS) with automated scaling for fluctuating traffic volumes. Built MLOps pipelines with MLflow and Airflow, automating experiment tracking, model versioning, and CI/CD deployments. Collaborated with data engineers to implement feature stores and distributed data transformations using PySpark and AWS Glue for petabyte-scale datasets. Designed network health dashboards in Grafana and Power BI, visualizing fault likelihood, latency trends, and customer churn risk. Introduced drift detection and model retraining triggers using statistical monitoring and CloudWatch alerts, maintaining high model precision over time. Partnered with data scientists to quantize and optimize deep learning models with ONNX Runtime, improving inference latency by 40%. Built customer segmentation models leveraging unsupervised learning (K-Means, DBSCAN) to classify high-risk churn segments for targeted retention campaigns. Integrated AWS Lambda for event-driven retraining workflows triggered by new data from the network operations center (NOC). Conducted A/B testing and simulation experiments to evaluate model-driven maintenance strategies, improving mean time to detect (MTTD) network faults by 19%. Worked with DevOps teams to enforce security, resilience, and performance compliance under AWS Well-Architected Framework. Deployed Graph Neural Network (GNN) prototypes on Neo4j and Amazon Neptune for root-cause analysis of interdependent network node failures. Designed real-time alerting and visualization dashboards for field engineers, integrating ML outputs into ServiceNow and internal monitoring tools. Collaborated closely with telecom product, NOC, and AI platform teams to align KPIs across model outputs, SLAs, and business goals. Environment: Python, TensorFlow, PyTorch, Scikit-learn, XGBoost, MLflow, FastAPI, Docker, Kubernetes (Amazon EKS), Apache Kafka, Spark Structured Streaming, Airflow, AWS SageMaker, AWS Lambda, AWS S3, AWS Glue, GCP Vertex AI, BigQuery, ONNX Runtime, Grafana, Prometheus, Power BI, Tableau, Git, GitHub Actions, Jenkins, Pandas, NumPy, spaCy, BERT, Neo4j, Amazon Neptune, SQL, NoSQL, PostgreSQL, MongoDB, Streamlit, Gradio, Linux, Bash, Agile, Jira, Confluence Client: UnitedHealth Group (UHG), North Carolina, USA Mar 2016 Oct 2019 Role: Data Scientist Project: Predictive Health Risk Stratification & Cost Optimization Platform Responsibilities: Developed predictive ML models (XGBoost, Random Forest, Logistic Regression) for forecasting readmissions, high-cost claimants, and chronic disease progression. Enhanced model accuracy through feature selection, correlation analysis, and cross-validation, achieving a ~25% improvement in prediction precision. Automated data preparation and ML workflows on AWS (S3, EC2, SageMaker, Lambda) for scalable training and deployment of healthcare models. Built end-to-end ML pipelines integrating data ingestion, transformation, model training, and inference orchestration through CI/CD frameworks. Implemented NLP-based analytics (spaCy, NLTK) to extract ICD/CPT codes, conditions, and treatment patterns from physician notes and unstructured EHR narratives. Prototyped deep learning models (TensorFlow 1.x, Keras) for sequential patient journey modeling and risk progression prediction. Partnered with data engineering teams to migrate legacy analytics workloads to cloud-native architectures on AWS and Databricks. Developed data visualization dashboards (Tableau, Power BI) to deliver patient risk trends, cost metrics, and care program effectiveness insights to stakeholders. Collaborated with clinical and actuarial teams to translate predictive insights into actionable care strategies and targeted intervention campaigns. Implemented model monitoring and drift detection pipelines to maintain accuracy and fairness over time using SHAP and internal validation frameworks. Integrated model governance and documentation processes aligning with HIPAA, PHI, and FDA audit compliance. Conducted A/B testing comparing ML-driven outreach vs. rule-based triggers, yielding a 17% higher early-risk detection rate. Mentored junior analysts and data engineers in feature engineering, ML evaluation, and cloud model deployment best practices. Presented insights and model outcomes to clinical leadership and operations teams, driving enterprise adoption of AI-based preventive care programs. Environment: Python, SQL, Pandas, NumPy, Scikit-learn, TensorFlow (v1.x), Keras, XGBoost, spaCy, NLTK, AWS (S3, EC2, SageMaker, Lambda), Tableau, Power BI, Git, Jupyter Notebook, CI/CD, Data Governance (HIPAA). Client: JP Morgan Chase, Jersey City, NJ December 2013 March 2016 Role: Python Developer/Data Engineer Project: Enterprise Risk & Customer Analytics Platform Responsibilities: Collaborated with the Data Engineering and Risk Analytics teams to design scalable data ingestion pipelines consolidating credit card, loan, and customer transaction data across multiple internal systems. Developed Python automation scripts for large-scale data extraction, cleansing, and transformation using requests, pandas, and numpy, improving processing efficiency by 30%. Engineered ETL workflows on Google Cloud Platform (GCP) using Cloud Dataflow, Pub/Sub, and BigQuery for real-time data streaming and aggregation from transactional sources. Built and optimized relational and dimensional data models (Star and Snowflake schema) in BigQuery and Teradata for marketing, credit, and compliance analytics. Created data validation frameworks in Python ensuring consistency across regulatory and operational data (Basel III, CCAR datasets). Implemented Apache Airflow DAGs for automated orchestration, dependency handling, and monitoring of daily data refresh pipelines. Partnered with the credit risk team to build machine-learning models (Logistic Regression, Decision Trees) predicting credit default and delinquency probabilities. Applied unsupervised learning (K-Means clustering) to segment customers by spending and repayment behavior for targeted offers. Designed and implemented fraud detection prototypes leveraging PySpark MLlib and transaction anomaly rules on streaming datasets. Developed and tuned complex SQL procedures, views, and materialized queries in Teradata to support downstream risk reporting. Built Tableau dashboards visualizing key banking KPIs delinquency trends, utilization ratios, churn indicators, and fraud risk scores. Containerized data pipelines using Docker for environment consistency across development, testing, and production. Collaborated with infrastructure teams to deploy ML models into GCP Cloud Functions and schedule retraining with Airflow. Supported data quality, lineage, and audit documentation to align with JP Morgan Chase s data governance and compliance standards. Conducted A/B testing to evaluate model effectiveness and fine-tuned hyperparameters for improved recall and precision in fraud detection. Integrated logging and monitoring via Stackdriver to ensure pipeline health and timely error alerting. Contributed to migration of on-prem ETL jobs to GCP, reducing data load time by 40% and operational cost by 25%. Presented analytics insights to business users and compliance teams to drive data-driven decisions on risk controls and marketing campaigns. Mentored junior developers on Python best practices, query optimization, and GCP data tooling. Environment: Python, PySpark, GCP (BigQuery, Dataflow, Pub/Sub, Cloud Functions, Stackdriver), SQL, Teradata, Apache Airflow, Docker, Tableau, Pandas, NumPy, Scikit-learn, BeautifulSoup, Matplotlib, Seaborn, AWS S3, Linux Keywords: cprogramm cplusplus continuous integration continuous deployment artificial intelligence machine learning javascript business intelligence sthree database rlang information technology California New Jersey