Senior Gen AI Engineer with Full-Stack Expertise

Pramod Kumar

Senior Gen AI Engineer

Phone +1-314-***-**** Email: ***********@*****.*** LinkedIn

PROFESSIONAL SUMMARY

Results-driven Full Stack & AI/ML Engineer with 8+ years of experience building Python-based APIs, React.js frontends, and cloud-native data platforms across AWS, Azure, and GCP, integrating advanced data engineering, MLOps, and LLM-based AI solutions.

Experienced in designing Retrieval-Augmented Generation (RAG) pipelines, autonomous agent frameworks, and multi-agent collaboration systems to accelerate Generative AI adoption in healthcare, finance, and telecom domains.

Proficient in Python, TensorFlow, PyTorch, and Scikit-learn, with deep hands-on experience fine-tuning LLMs (GPT, LLaMA, Falcon) for text generation, reasoning, and task planning.

Implemented agentic frameworks such as LangChain, LangGraph, CrewAI, AutoGen, and LlamaIndex to enable agent-to-agent communication, dynamic task execution, and autonomous workflows.

Designed and deployed RAG systems leveraging vector databases like Pinecone, FAISS, and ChromaDB to ground LLMs in domain-specific knowledge, improving accuracy and explainability.

Fine-tuned and customized foundation models using prompt engineering, LoRA, and adapter-based MCP techniques to align LLM behavior with enterprise objectives.

Integrated AI agents with Power BI, Tableau, and Looker Studio to provide explainable, real-time insights for business users.

Architected and deployed cloud-native data warehouses and data lakes on Snowflake, Redshift, BigQuery, and Azure Synapse, supporting predictive analytics and ML model integration.

Built ETL/ELT and streaming frameworks using AWS Glue, Azure Data Factory, GCP Dataflow, PySpark, Kafka, and Kinesis, enabling batch and real-time AI data flows.

Hands-on expertise with PostgreSQL (Aurora, RDS, EDB Postgres) including logical & physical replication, PITR, query tuning, indexing, and pgBackRest for enterprise-grade HA/DR.

Developed Python microservices and FastAPI endpoints to expose AI/ML models as production-ready APIs integrated with enterprise and SAP applications.

Automated end-to-end ML pipelines using Airflow (MWAA / Cloud Composer), AWS Step Functions, and Azure Logic Apps, with CI/CD integration via Terraform, Jenkins, GitLab CI/CD, and GitHub Actions.

Deployed scalable GenAI workloads on Vertex AI, Azure ML Studio, and AWS SageMaker, optimizing inference latency, cost efficiency, and governance.

Applied SRE principles (SLOs, SLIs, SLAs) and AI-driven auto-remediation workflows to ensure reliability and operational resilience across AI pipelines.

Embedded observability and governance controls using OpenTelemetry, Prometheus, Grafana, Lake Formation, Purview, and Dataplex, ensuring compliance with HIPAA, SOC2, and internal security standards.

Managed container orchestration using Kubernetes (AKS, GKE) and Helm, ensuring secure, scalable, and automated AI deployments in multi-cloud environments.

Designed and implemented Master Data Management (MDM) frameworks using Reltio, Semarchy, and Informatica MDM to centralize Party, Provider, and Clinical domains.

Integrated MDM solutions with BigQuery, Dataplex, and transactional systems for cross-domain data consistency, lineage tracking, and real-time synchronization.

Partnered with Data Science, SRE, and DevOps teams to define RTO/RPO objectives, create operational runbooks, and ensure system reliability in production AI environments.

Recognized as a self-motivated leader and mentor with a track record of delivering enterprise-grade AI/ML solutions, ensuring data quality, security, and regulatory compliance, while driving innovation in multi-cloud, RAG, and agentic AI systems.

TECHNICAL SKILLS

Databases:

PostgreSQL (Aurora, RDS, EDB Postgres, PGD – Postgres Distributed), SQL Server, Oracle, MySQL, Teradata, Azure SQL Database, Snowflake, NoSQL (DynamoDB, MongoDB, Cassandra, HBase).

PostgreSQL Operations:

Replication (logical & physical), Multi-AZ configuration, Read Replicas, Failover, PITR (Point-in-Time Recovery), Parameter Groups, pgBouncer, pgBackRest, Autovacuum tuning, EXPLAIN/ANALYZE, Indexing (BTREE, GIN, GiST), Partitioning, and Performance Tuning.

Cloud Platforms:

AWS: Glue, Lambda, Step Functions, Lake Formation, S3, Redshift, EMR, Kinesis, SageMaker.

Azure: Data Factory, Synapse, Azure ML, Cognitive Services, Azure OpenAI.

GCP: BigQuery, Vertex AI, Pub/Sub, Dataflow, Dataproc.

AI/ML & Agentic Frameworks:

Python (pandas, NumPy, SciPy, scikit-learn, TensorFlow, PyTorch), LangChain, LlamaIndex, AutoGen, Hugging Face Transformers, Retrieval-Augmented Generation (RAG), Pinecone, FAISS, and ChromaDB.

Data Engineering & Big Data:

Spark (PySpark, Scala, Databricks, EMR), Kafka (Connect, Streams), Hadoop Ecosystem (HDFS, YARN, Hive, Pig, Sqoop, Oozie, Flume, NiFi, Flink, MapReduce), and ETL/ELT tools including SSIS, Talend, DataStage, and Alteryx.

Programming & Scripting:

Python, SQL, Scala, R, Bash, Shell, T-SQL, PL/SQL, HiveQL, C/C++, and JavaScript.

Automation, DevOps & MLOps:

Terraform, Ansible, Jenkins, Git, GitHub, GitLab CI/CD, Bitbucket, Docker, Kubernetes (EKS, GKE, AKS), Apache Airflow (MWAA, Cloud Composer), AWS Step Functions, and ML pipeline automation.

Governance & Metadata Management:

AWS Lake Formation, Microsoft Purview, Informatica, Dataplex, Data Catalogs, Metadata and Lineage Tracking, Data Quality Rules, and IAM/Role-Based Access Controls.

Monitoring & Reporting:

CloudWatch, CloudTrail, Datadog, Prometheus, Grafana, pganalyze, Splunk, Power BI, Tableau, QuickSight, Looker Studio, and Stackdriver.

Other Tools & Methodologies:

Jira, Slack, Postman, Microsoft Excel, Agile, Scrum, Waterfall, Test-Driven Development (TDD), and CI/CD best practices.

PROFESSIONAL EXPERIENCE

JPMorgan Chase – New York, NY

Senior Gen AI Engineer Jun 2023 – Present

Responsibilities:

Enterprise AI & Data Architecture: Designed and deployed multi-cloud data and AI platforms across AWS and GCP (Vertex AI) integrating PostgreSQL (Aurora/RDS), Redshift, and Snowflake with LLM-based GenAI services. Architected data pipelines that support Retrieval-Augmented Generation (RAG) for internal knowledge search, report summarization, and compliance automation.

Database Architecture & Administration: Engineered fault-tolerant PostgreSQL clusters using Aurora and RDS with Multi-AZ replication, read replicas, PITR, and cross-region failover. Modernized legacy Oracle/MS SQL Server workloads into Aurora PostgreSQL, ensuring zero-loss migrations for high-volume financial applications and enabling downstream integration with AI analytics layers.

LLM & RAG Integration: Built and deployed RAG pipelines combining LangChain, LlamaIndex, and Pinecone for context-aware document retrieval. Integrated CrewAI and AutoGen to orchestrate multi-agent collaboration for automating compliance report generation, code review, and data-quality anomaly summarization.

Performance Optimization & Automation: Tuned high-transaction workloads using EXPLAIN/ANALYZE, advanced indexing (BTREE, GIN, GiST), and table partitioning, improving query efficiency by 30%. Automated backup, failover, and scaling using pgBackRest, pgBouncer, and Terraform, with end-to-end CI/CD through Jenkins and GitLab.

Data Engineering Pipelines: Built cloud-native ETL and ELT frameworks using AWS Glue, Spark (PySpark), Databricks, and Kinesis, ingesting structured and unstructured data into Aurora, Redshift, and Vertex AI Feature Store. Enabled streaming analytics for fraud detection, risk modeling, and financial forecasting.

Generative AI Application Layer: Developed FastAPI microservices that expose LLM inference endpoints integrating OpenAI, Vertex AI, and Azure OpenAI APIs. Supported dynamic prompt-engineering, memory caching, and semantic search modules for internal developer tools and business analytics dashboards.

MLOps / DataOps Orchestration: Implemented end-to-end AI pipelines using Airflow (MWAA) and AWS Step Functions, with model versioning, retraining, and evaluation managed through MLflow and GitHub Actions. Established CI/CD workflows for deploying RAG agents, ensuring governance, lineage, and reproducibility across dev, test, and prod.

Observability & Responsible AI: Established monitoring dashboards using CloudWatch, Prometheus, Grafana, and Datadog to track LLM token usage, latency, and accuracy metrics. Integrated OpenTelemetry traces with LangSmith for prompt-performance evaluation and fine-tuning recommendations.

Big Data & Streaming: Engineered real-time ingestion and lakehouse architecture using AWS Kinesis, Iceberg, and S3 to deliver schema-evolution-ready, low-latency datasets for GenAI workloads. Implemented event-driven ingestion patterns with Kafka and Pub/Sub for downstream model consumption in Vertex AI Pipelines.

Security & Compliance: Enforced cloud security controls using IAM, KMS, TLS encryption, and RBAC. Configured logging and audit policies ensuring compliance with SOX, GDPR, and internal financial data-governance standards for all GenAI and data pipelines.

Cross-Functional Leadership: Partnered with SRE, MLOps, and Data Science teams to design resilient AI systems meeting RTO/RPO requirements. Authored detailed architecture diagrams, runbooks, and knowledge-base documentation, mentoring junior engineers on RAG, MLOps, and PostgreSQL optimization best practices.

Environment: AWS (Aurora PostgreSQL, Redshift, Glue, EMR, Kinesis, Lambda, S3, SageMaker, Step Functions), GCP (Vertex AI, BigQuery, Dataflow, Pub/Sub, Dataproc), Azure (ADF, Synapse, Azure ML, OpenAI), Databricks, Spark (PySpark/Scala), Kafka, Terraform, Jenkins, GitLab, MLflow, LangChain, LangGraph, CrewAI, AutoGen, Pinecone, FAISS, Prometheus, Grafana, CloudWatch, Datadog, Tableau, Looker, Linux/Unix.

State Farm Insurance – Bloomington, IL

AI/ML Software Engineer Mar 2021 – May 2023

Responsibilities:

Database Administration: AI-Driven Data Modernization: Led end-to-end modernization of legacy Oracle and SQL Server workloads into AWS Aurora PostgreSQL and GCP BigQuery, enabling AI-ready datasets for downstream Generative AI use cases such as policy summarization, claim triage, and document intelligence.

Database Engineering & Performance: Administered and optimized PostgreSQL (Aurora, RDS) and SQL Server databases, implementing indexing, query tuning, and schema refactoring that improved data retrieval speeds by 25%. Designed and automated HA/DR architectures using pgBackRest, read replicas, and PITR, ensuring continuous availability for mission-critical insurance systems.

AI & RAG Pipeline Enablement: Collaborated with data science teams to prepare and structure data for RAG (Retrieval-Augmented Generation) applications. Designed pipelines that integrated LangChain and LlamaIndex to fetch contextual policy information and claims data from PostgreSQL, S3, and Snowflake, improving model grounding and response accuracy for internal LLM copilots.

NoSQL & Distributed Systems: Designed and deployed Cassandra and DynamoDB clusters on AWS, tuning replication and consistency levels for high-volume claim processing workloads. Integrated these with AI feature stores to support hybrid transactional/analytical use cases for fraud detection and customer analytics.

Data Engineering & Streaming: Built real-time ingestion pipelines using Kafka, PySpark, and AWS EMR, integrating data streams from claims and telematics systems into Aurora and Redshift. Enabled real-time model inference and automated decision-making workflows for risk scoring and underwriting.

MLOps & Automation: Automated end-to-end ML lifecycle using Airflow, Step Functions, and Terraform, integrating with CI/CD pipelines built in Jenkins and GitLab. Containerized ML workloads using Docker and deployed predictive models and AI agents on Vertex AI and SageMaker, ensuring reproducibility and fast iteration cycles.

Agentic AI Frameworks: Supported pilot projects exploring agentic AI orchestration with CrewAI, LangGraph, and AutoGen, enabling collaborative LLM agents to perform claims document classification, sentiment extraction, and call-center summarization.

Observability & Responsible AI: Implemented real-time observability for AI data pipelines using CloudWatch, Datadog, and Splunk, capturing token usage, latency, and model performance metrics. Established traceability and governance dashboards using OpenTelemetry and Prometheus for proactive model evaluation.

Analytics & Visualization: Delivered business insights through Hive, Impala, and PySpark queries visualized in Power BI and Tableau, supporting actuarial analytics, premium optimization, and customer engagement metrics.

Security & Compliance: Applied enterprise security standards across all data and AI pipelines—implemented Kerberos authentication, IAM role-based policies, and KMS encryption to meet HIPAA, SOX, and GDPR compliance for sensitive insurance data.

Collaboration & Knowledge Sharing: Partnered with SRE, application development, and MLOps teams to design AI-ready, fault-tolerant architectures. Authored operational runbooks, data dictionaries, and AI deployment guides, promoting cross-team alignment and best practices in DataOps, RAG, and AI observability.

Environment: AWS (Aurora PostgreSQL, Redshift, EMR, Glue, S3, Lambda, Step Functions), GCP (Vertex AI, BigQuery, Dataflow, Pub/Sub), Databricks, Spark (PySpark/Scala), Kafka, Cassandra, DynamoDB, Terraform, Ansible, Jenkins, GitLab, Airflow, Oozie, Hive, Sqoop, Docker, Power BI, Tableau, Prometheus, CloudWatch, Datadog, Splunk, Linux/Unix.

Cardinal Health – Dublin, OH

Data Engineer Jan 2020 – Feb 2021

Responsibilities:

Enterprise Data Engineering & AI Integration: Designed and implemented end-to-end multi-cloud ETL and streaming pipelines across AWS, GCP, and on-prem Hadoop to support advanced AI/ML analytics for healthcare data. Enabled near-real-time ingestion and transformation of clinical, pharmacy, and logistics datasets, preparing high-quality, compliant data for use in Generative AI and predictive modeling workflows.

AI-Ready ETL Architecture: Built hybrid batch + streaming pipelines using Kafka, Pub/Sub, Spark, PySpark, Dataflow, and Databricks, integrating structured and semi-structured healthcare data into Aurora PostgreSQL, Redshift, and BigQuery. Implemented DataOps automation for lineage, quality, and schema enforcement, achieving a 40% reduction in ETL latency and a 25% boost in downstream analytics throughput.

PostgreSQL Administration & Modernization: Administered and optimized Aurora PostgreSQL (RDS) clusters, implementing logical/physical replication, PITR, query optimization, and index tuning for mission-critical healthcare applications. Migrated legacy Oracle and SQL Server workloads to Aurora with zero data loss and improved query performance by over 30%.

EDB Postgres & PGD Evaluation: Evaluated EnterpriseDB Postgres and Postgres Distributed (PGD) features to design high-availability and geo-distributed replication strategies suitable for HIPAA-regulated healthcare data ecosystems. Developed resilient HA/DR blueprints incorporating RTO/RPO objectives and cross-region failover simulations.

RAG & Generative AI Enablement: Partnered with AI teams to design Retrieval-Augmented Generation (RAG) pipelines using LangChain, LlamaIndex, and Pinecone, enabling clinical documentation search and summarization use cases. Integrated FAISS and ChromaDB vector indexes to allow LLMs to retrieve de-identified patient information securely, improving accuracy and contextual grounding of AI responses.

Healthcare Data Integration: Consolidated NoSQL and Hadoop data sources (Cassandra, Hive) into relational systems (Aurora, Redshift, Snowflake) to provide unified, analytics-ready data for predictive modeling, forecasting, and regulatory reporting. Developed reusable Spark transformations supporting ML model training and batch inference pipelines.

Data Governance & Compliance: Implemented metadata and lineage frameworks inspired by Informatica and Microsoft Purview, ensuring data traceability and policy enforcement across healthcare domains. Applied data masking, tokenization, and access controls to maintain compliance with HIPAA, SOC2, and internal audit standards.

Monitoring & Observability: Developed end-to-end observability dashboards using ELK, Datadog, and CloudWatch, tracking pipeline throughput, replication lag, and query latency. Integrated alerting and anomaly detection workflows, reducing mean time to identify (MTTI) incidents by 35%.

MLOps & Automation: Automated infrastructure provisioning and pipeline deployments using Terraform, Ansible, and Jenkins, achieving consistent environment setups across dev, QA, and prod. Supported ML model integration into data pipelines through Airflow DAGs and containerized services, accelerating model deployment lifecycles by 20%.

Analytics & Visualization: Developed SQL-based analytics layers and Power BI/Tableau dashboards for performance monitoring, patient demand forecasting, and supply-chain optimization. Provided data to feed AI-driven forecasting and resource allocation models, reducing manual reporting efforts.

Collaboration & Documentation: Partnered with cross-functional Data Science, DevOps, and Security teams to design AI-compatible, resilient data infrastructures. Authored detailed runbooks, architecture diagrams, and best-practice guides, mentoring junior engineers on database administration, ETL orchestration, and AI data readiness.

Environment: AWS (Aurora PostgreSQL, Redshift, S3, Glue, Lambda, EMR), GCP (BigQuery, Dataflow, Pub/Sub, Vertex AI), Hadoop (Hive, Sqoop, Hortonworks), Spark (PySpark, SparkSQL), Databricks, Kafka, Cassandra, Snowflake, EDB Postgres, Terraform, Ansible, Jenkins, ELK, Datadog, Power BI, Tableau, Unix/Linux.

Krios Info Solutions – Hyderabad, India

ETL Developer Jul 2017 – Oct 2019

Responsibilities:

Big Data Ecosystem Development: Designed and implemented large-scale data integration frameworks using Hadoop, Spark, SparkSQL, and HBase, enabling enterprise data teams to perform faster analytics and high-volume data processing across distributed clusters. Contributed to scalable data ingestion architectures that powered downstream business intelligence and machine learning initiatives.

ETL & Data Pipeline Automation: Developed and optimized ETL workflows using Hive, Pig, and Sqoop, automating ingestion from Oracle, SQL Server, and MySQL into HDFS and Hive warehouses. Improved data processing throughput by 30% through SQL tuning and Spark job optimization, reducing end-to-end data latency in reporting pipelines.

Cluster & Infrastructure Management: Monitored and maintained multi-node Hadoop clusters, performing node scaling, health checks, log analysis, and fault recovery procedures. Automated cluster configuration and node provisioning scripts to ensure optimal resource utilization and high system availability.

Data Migration & Transformation: Supported enterprise database migrations and data normalization projects, leveraging Sqoop, Hive, and SparkSQL to move data between on-prem relational systems and Hadoop Distributed File System (HDFS). Improved schema consistency and optimized partitioning strategies for downstream analytical workloads.

Custom Application Development: Designed Java-based client utilities and RESTful web services to automate ingestion of large datasets from Oracle and SQL Server into HDFS, ensuring seamless connectivity and integration with upstream data sources. Enhanced data validation and logging mechanisms for traceability and accuracy.

Workflow Orchestration & Scheduling: Automated ETL job execution and dependency tracking using Oozie, integrating Hive, Sqoop, and Spark pipelines into unified workflows. Reduced manual interventions by implementing job alerts, checkpoints, and retry mechanisms for high-volume nightly loads.

Data Quality & Test-Driven Development: Applied Test-Driven Development (TDD) principles to build reusable data validation and reconciliation scripts, embedding them within CI/CD workflows. Increased data accuracy and reliability through automated regression checks before production releases.

Hybrid Cloud & Early Cloud Adoption: Participated in hybrid cloud enablement projects using Microsoft Azure, assisting in migration design sessions and proof-of-concept deployments for Hadoop workloads. Gained early exposure to Azure Data Factory, Blob Storage, and Synapse Analytics integration for cloud-based processing.

Collaboration & Agile Practices: Worked in a Scrum-based Agile environment, contributing to sprint planning, backlog refinement, and cross-functional coordination. Delivered iterative data solutions aligned with evolving business and analytics requirements.

Mentorship & Documentation: Authored ETL runbooks, cluster operations guides, and data workflow documentation to streamline maintenance and onboarding. Supported junior developers in adopting Spark-based transformations and adhering to coding standards.

Environment: Hadoop (HDFS, YARN, MapReduce, Hive, Pig, Sqoop, Oozie, HBase), Spark (PySpark, SparkSQL, Databricks), Oracle 10g, SQL Server, MySQL, Microsoft Azure (Data Factory, Blob Storage, Synapse), Java, Linux/Unix, Git, Agile/Scrum.

Education:

Bachelor in computer Science from Loyola Academy Degree & PG college

Contact this candidate