Home

Tandava Krishna Doddi - Sr AI Ml Engineer/ Data Engineer / Gen AI Engineer

[email protected]

Location: Remote, Remote, USA

Relocation: Yes

Visa: H1b

PROFESSIONAL SUMMARY Senior AI/ML Engineer and GenAI Architect with 10+ years of end-to-end experience designing, developing, and deploying enterprise-grade AI/ML solutions, LLM-powered applications, agentic workflows, and production-ready data engineering platforms across AWS, Azure, and GCP. Deep hands-on expertise in Generative AI including LLM orchestration (LangChain, LangGraph, AutoGen, CrewAI), Retrieval-Augmented Generation (RAG), prompt engineering (design, optimization, context management), and multi-step agent-based automation frameworks targeting enterprise decision support and workflow automation. Proven track record operationalizing AI in production: end-to-end MLOps lifecycle covering model training, experimentation (MLflow), fine-tuning approaches (LoRA, QLoRA, PEFT), pipeline orchestration (Kubeflow, DVC), model serving, drift monitoring, and LLMOps observability for GenAI systems in regulated financial and healthcare environments. Extensive experience building and integrating AWS AI services Amazon Bedrock (Claude, Titan, Llama), SageMaker, Lambda, Glue, Kinesis, S3, and CloudWatch into scalable, event-driven GenAI and ML workflows delivering measurable business outcomes. Strong expertise in RAG architectures and vector search infrastructure: Pinecone, FAISS, pgvector, Weaviate, and Chroma for embeddings-based semantic retrieval, hybrid search, and domain-specific intelligence extraction from structured and unstructured enterprise data sources. Expert Python developer for data ingestion, feature engineering, model inference services, and RESTful API design; proficient in containerized microservices (Docker, Kubernetes) and infrastructure-as-code (Terraform) for scalable, reproducible AI deployments. Advanced data engineering capabilities: designing high-throughput batch and real-time data pipelines using Apache Spark, Kafka, Hadoop, Snowflake, and Databricks to power ML model training, feature stores, and analytics platforms at enterprise scale. Strong communicator and cross-functional collaborator experienced translating complex AI/ML concepts into actionable technical roadmaps for data engineers, software developers, data scientists, and business stakeholders in Agile delivery environments. TECHNICAL SKILLS Programming Python (Advanced), SQL, Java GenAI & LLMs LangChain, LangGraph, AutoGen, CrewAI, DSPy, OpenAI API, Amazon Bedrock (Claude, Titan, Llama), Hugging Face Transformers, Prompt Engineering, RAG Architecture, Agent Orchestration, Agentic Workflows ML Frameworks PyTorch, TensorFlow, scikit-learn, PySpark, Pandas, NumPy, Jupyter Notebook, MLflow, SageMaker LLM Fine-Tuning LoRA, QLoRA, PEFT, Model Evaluation & Benchmarking, Explainable AI (SHAP, LIME), Embeddings, Tokenization, Context Window Management Vector & Search Pinecone, FAISS, pgvector, Weaviate, Chroma, Semantic Search, Hybrid Retrieval, Chunking Strategies, Embedding Pipelines MLOps & LLMOps MLflow, Kubeflow, DVC, Model Registry, CI/CD for ML (GitHub Actions, Jenkins), Docker, Kubernetes, Terraform, Model Drift Monitoring, LLMOps Observability Cloud Platforms AWS (Bedrock, SageMaker, Lambda, S3, Glue, Kinesis, IAM, CloudWatch, EC2, Step Functions, Snowpipe), Azure (AML, DP-203), GCP (Vertex AI exposure) Data Engineering Apache Spark, Apache Kafka, Kafka Streams, ETL/ELT Pipelines (Batch & Real-Time), Apache NiFi, Hadoop (HDFS, Hive), Snowflake, Databricks, Delta Lake, dbt Databases & Storage PostgreSQL, SQL Server, Oracle, MySQL, MongoDB, Snowflake, Redshift API & Microservices RESTful API Design, FastAPI, Flask, Event-Driven Architecture, Kafka-based Integration, API Gateway Observability ELK Stack, Grafana, CloudWatch, Prometheus, Logging & Alerting, Model Performance Monitoring, Pipeline SLA Tracking Governance & Security Data Governance, RBAC, Encryption, Audit Logging, GDPR/HIPAA Compliance, Responsible AI, Prompt Injection Defense Methodologies Agile/Scrum, SDLC, Microservices Architecture, Cross-functional Collaboration, Technical Roadmapping PROFESSIONAL EXPERIENCE Cardinal Health | Dublin, OH | Dec 2024 Present Senior AI/ML Engineer | GenAI Architect Architected and deployed enterprise-grade GenAI solutions using Amazon Bedrock (Claude 3, Titan, Llama models) and LangChain/LangGraph, enabling automated data enrichment, entity resolution, and intelligent decision-support workflows across multi-terabyte healthcare data environments. Designed and implemented end-to-end Retrieval-Augmented Generation (RAG) pipelines integrating FAISS and pgvector vector stores, embedding generation (Hugging Face, Bedrock embeddings), document chunking strategies, and semantic retrieval layers enabling domain-specific intelligence extraction from large unstructured healthcare document repositories. Built multi-step LLM agent workflows using LangChain and LangGraph for agentic task orchestration including tool-calling agents, reasoning chains, and memory-enabled conversational pipelines reducing manual data review effort by automating classification and routing decisions. Developed advanced prompt engineering strategies: prompt design, few-shot optimization, chain-of-thought structuring, and context window management to maximize LLM accuracy, reliability, and cost-efficiency in production healthcare data workflows. Engineered ML-based entity resolution models using Python (PyTorch, scikit-learn) and statistical similarity techniques (cosine similarity, Jaro-Winkler, fuzzy matching) to deduplicate and reconcile patient and supplier records across disparate ERP systems, improving master data quality by over 30%. Designed and implemented anomaly detection pipelines using isolation forests, autoencoders, and statistical control charts to identify data quality anomalies, processing irregularities, and operational outliers across batch and near-real-time healthcare data streams. Built scalable feature engineering pipelines on AWS Glue and Lambda, processing structured and semi-structured datasets (JSON, Parquet, CSV) from S3 data lake into ML-ready feature sets for model training, validation, and real-time inference serving. Operationalized full MLOps lifecycle: MLflow for experiment tracking, model versioning, and model registry; CI/CD pipelines (GitHub Actions, Jenkins) for automated model testing, containerized deployment (Docker), and staged production rollouts with rollback capability. Implemented LLMOps observability stack: end-to-end logging of prompt/response pairs, token usage tracking, latency monitoring, hallucination detection heuristics, and drift alerting to maintain production reliability and performance baselines for Bedrock-based GenAI services. Integrated GenAI capabilities into backend microservices via RESTful APIs (FastAPI), enabling seamless enterprise workflow automation where LLM-powered services are embedded into existing data processing and approval workflows. Applied Explainable AI techniques (SHAP, LIME) to interpret anomaly detection and classification model outputs, improving transparency and business trust in ML-driven decisions across compliance-sensitive healthcare operations. Implemented enterprise-grade data security controls: RBAC, field-level encryption, audit logging, and compliance guardrails (HIPAA-aligned) across all ML pipelines and GenAI service endpoints. Collaborated with product owners, data engineers, compliance officers, and business analysts in Agile/Scrum ceremonies to translate healthcare business objectives into actionable AI/ML technical roadmaps and sprint deliverables. Conducted model performance reviews, root cause analysis for pipeline failures, and iterative improvement cycles reducing data pipeline error rates and improving model inference throughput by over 25% through query optimization and batching strategies. Tech Stack: Python, LangChain, LangGraph, Amazon Bedrock (Claude/Titan/Llama), FAISS, pgvector, RAG, Prompt Engineering, Agent Orchestration, PyTorch, scikit-learn, MLflow, FastAPI, Docker, Kubernetes, AWS (Glue, Lambda, S3, Kinesis, CloudWatch, Bedrock, Step Functions), Apache Kafka, CI/CD (GitHub Actions, Jenkins), SHAP/LIME, RBAC, Agile/Scrum USAA | San Antonio, TX | Dec 2022 Nov 2024 Senior Data Engineer | ML Engineer Designed and deployed scalable ML-powered analytics infrastructure supporting real-time fraud detection, anomaly detection, and risk scoring across high-volume financial transaction data processing over 50 million daily events across USAA's transactional and clickstream data estate. Built real-time ML inference and feature serving pipelines using Kafka Streams and Apache Flink for low-latency event-driven decision-making, achieving sub-100ms inference latency for fraud scoring models deployed in production payment processing flows. Developed and productionized anomaly detection models using Python (scikit-learn, PySpark) including Isolation Forest, Local Outlier Factor, and LSTM-based sequential anomaly detection to identify fraudulent transaction patterns and operational risk signals in streaming financial data. Implemented feature engineering and data transformation workflows integrated with Hadoop (HDFS, Hive) and Cloudera CDP, building reusable feature pipelines that served multiple downstream ML models with versioned, lineage-tracked feature datasets. Designed end-to-end ML experimentation workflows in Jupyter Notebook and MLflow environments covering hypothesis definition, feature selection, hyperparameter tuning (Optuna, GridSearch), cross-validation, and A/B model evaluation enabling data-science teams to accelerate model iteration cycles. Built high-throughput batch data ingestion pipelines using Apache NiFi and Spark to ingest and transform structured and semi-structured financial data (JSON, Avro, Parquet) from upstream APIs, SFTP sources, and message queues into Hive-backed analytical data stores. Implemented a comprehensive monitoring and observability framework using ELK Stack (Elasticsearch, Logstash, Kibana) and Grafana dashboards to track pipeline performance, data quality SLAs, model prediction accuracy, and infrastructure health across ML systems in production. Operationalized CI/CD for ML pipelines using Git-based workflows, enabling automated unit testing, data validation checks, model regression testing, and blue-green deployments for ML model updates with zero downtime. Applied statistical modeling techniques (logistic regression, time-series decomposition, survival analysis) to identify behavioral trends, seasonal patterns, and risk signals across customer transaction histories, informing downstream business rules and alert thresholds. Enforced financial-grade data governance and compliance: RBAC, LDAP-based access controls, column-level data masking, data lineage tracking, and SOX-aligned audit logging across regulated financial data environments. Built data quality and validation frameworks using Great Expectations to enforce schema validation, null checks, distribution tests, and referential integrity constraints across ML feature pipelines preventing bad data from propagating to model training or scoring jobs. Collaborated with data scientists, software engineers, risk analysts, and business stakeholders to define ML model requirements, validate model outputs against business logic, and deliver production-grade ML solutions aligned with USAA's financial risk and compliance standards. Provided L3 production support: triaged and resolved data pipeline incidents, diagnosed model performance degradation, optimized slow Spark jobs (partition tuning, broadcast joins, caching strategies), and resolved Kafka consumer lag issues under SLA constraints. Tech Stack: Python, scikit-learn, PySpark, MLflow, Kafka Streams, Apache Kafka, Apache NiFi, Hadoop (HDFS, Hive), Cloudera CDP, Jupyter Notebook, Great Expectations, ELK Stack, Grafana, CI/CD (Git), Optuna, MongoDB, PostgreSQL, Oracle, RBAC/LDAP, SOX Compliance, Agile/Scrum Best Buy | Minneapolis, MN | Apr 2021 Nov 2022 Data Engineer | ML Engineer Translated retail business requirements into scalable ML-ready data architectures, designing end-to-end data pipelines that ingested, transformed, and served product catalog, inventory, and customer transaction data to support machine learning, analytics, and personalization use cases. Built and optimized ETL pipelines using Python, SQL, and Informatica PowerCenter to process high-volume structured and semi-structured retail data integrating POS systems, e-commerce platforms, and third-party vendor feeds into a centralized analytical data store. Designed and maintained feature engineering pipelines to produce ML-ready datasets for downstream recommendation models, demand forecasting models, and customer segmentation algorithms including feature normalization, encoding, and aggregation logic at scale using Apache Spark. Applied Python-based statistical analysis and anomaly detection techniques (Z-score, IQR, isolation forest) to identify data quality issues, inventory discrepancies, pricing anomalies, and sales pattern outliers across retail operational datasets. Developed complex SQL transformations and PySpark data processing logic for high-performance batch pipelines, including window functions, slowly changing dimensions (SCD Type 2), and incremental load patterns to support large-scale retail data warehouse operations. Designed and validated ML data workflows in Jupyter Notebook environments, building reusable preprocessing and validation modules enabling data scientists to rapidly prototype and iterate on new models with clean, well-documented datasets. Integrated data pipelines with downstream BI and analytics platforms (Power BI, Tableau) and ML inference endpoints enabling real-time and batch data availability for KPI reporting, predictive analytics dashboards, and business decision support tools. Implemented CI/CD practices (Jenkins, Git) for data pipeline deployment and validation including automated schema checks, row count validations, and data freshness assertions reducing data incident rates and improving pipeline release reliability. Monitored pipeline performance using custom alerting and logging solutions; conducted bottleneck analysis and implemented Spark optimization strategies (partition pruning, predicate pushdown, executor tuning) to improve job throughput by approximately 35%. Implemented RBAC, audit logging, and data masking standards across retail data pipelines, ensuring PII protection and compliance with enterprise data governance policies. Contributed to Agile/Scrum delivery cycles, participating in sprint planning, backlog grooming, technical design reviews, and cross-team data architecture discussions. Tech Stack: Python, SQL, PySpark, Jupyter Notebook, scikit-learn, ETL Pipelines, Informatica PowerCenter, Apache Spark, AWS (S3, Glue, basic), CI/CD (Jenkins/Git), Power BI, Tableau, SQL Server, Oracle, Teradata, RBAC, Audit Logging, Agile/Scrum ADP | Hyderabad, India | May 2017 Dec 2020 Data Engineer Designed, built, and maintained ETL pipelines and data processing workflows supporting business intelligence, payroll analytics, and HR operational reporting across ADP's enterprise data platform. Performed time-series analysis and trend modeling using Python (Pandas, NumPy) on asset and resource utilization datasets, identifying variance patterns that contributed to a ~70% improvement in project completion forecasting accuracy. Developed KPI dashboards and executive visualizations in Tableau covering resource utilization, net profit margin, gross profit margin, and operational burn rate enabling real-time performance monitoring for senior business stakeholders. Wrote advanced SQL queries leveraging analytic functions window functions, cumulative distribution, NTILE, RANK, LAG/LEAD for complex data validation, audit reconciliation, and multi-dimensional reporting across payroll and HR data marts. Collaborated with ETL developers, data architects, and project managers on data mapping, migration planning, and application architecture design; produced Requirements Traceability Matrices (RTMs) and Root Cause Analysis (RCA) reports for production data incidents. Recommended data-driven solutions to increase revenue, reduce operational expenses, and maximize compliance and data quality across business units, translating analytical findings into executive-level presentations and action plans. Conducted data profiling, cleansing, and preprocessing using advanced Excel (Pivot Tables, VLOOKUP, INDEX/MATCH, conditional logic) and SQL validation scripts to ensure data accuracy and integrity of source-to-target loads. Tech Stack: Python, Pandas, NumPy, SQL, Tableau, ETL, KPI Dashboards, RTMs, RCA, Advanced Excel, Data Validation, SSIS (exposure), Oracle, SQL Server HCL Technologies | Hyderabad, India | Jun 2014 Apr 2017 Data Analyst | ETL Developer Designed, developed, and deployed end-to-end ETL processes using SSIS and Java to extract, transform, and load data from heterogeneous source systems (flat files, APIs, RDBMS) into SQL Server data warehouses and data marts, ensuring data consistency and processing reliability. Built and optimized SSIS packages incorporating complex data flow transformations, conditional split logic, derived columns, fuzzy lookups, and error-handling mechanisms; deployed and automated execution via SQL Server Agent with dependency-aware scheduling. Implemented data validation and cleansing logic within SSIS pipelines using SQL Server Data Quality Services (DQS) and custom Script Tasks/Transformations, reducing downstream data quality defects and improving master data consistency across the warehouse. Developed SSRS reporting solutions with complex parameterization, drill-down capabilities, conditional formatting, and scheduled delivery mechanisms; collaborated with business stakeholders to gather requirements and translate them into high-quality analytical reports. Performed SSIS performance tuning buffer size optimization, parallel data flow execution, index-aligned staging strategies, and incremental load redesign to meet data processing SLAs for overnight ETL runs serving enterprise reporting consumers. Designed and documented data models, source-to-target mapping documents, and ETL technical specifications in coordination with data architects and business analysts, ensuring traceability and maintainability of warehouse solutions. Tech Stack: SQL Server, SSIS, SSRS, ETL, Java, T-SQL, Data Warehouse, Data Mart, SQL Server Agent, DQS, Data Modeling Keywords: continuous integration continuous deployment artificial intelligence machine learning business intelligence sthree Minnesota Ohio Texas