AI & GenAI Engineer with MLOps Expertise

Vinay Gajevalli

+1-314-***-**** ***********************@*****.*** www.linkedin.com/in/vinay-g-b593a2360 Decatue, IL

Professional Summary

AI & GenAI Engineer with almost 5 years of end-to-end experience building production-grade machine learning, LLM, and multi-agent systems across banking, healthcare, and enterprise environments, with a proven ability to take solutions from prototype to scalable deployment.

Highly skilled in multi-agent AI architecture, leveraging LangGraph, AutoGen, CrewAI, and LangChain to design coordinated agent ecosystems capable of task delegation, contextual reasoning, and dynamic decision-making with optional human-in-the-loop controls.

Deep expertise in Retrieval-Augmented Generation (RAG), including vector database design, embedding pipelines, and retrieval optimization using FAISS, Pinecone, Weaviate, OpenSearch, and ChromaDB to enable low-latency, context-aware LLM responses.

Experienced in customizing and optimizing LLM behavior, implementing reinforcement learning techniques such as RLHF and RLAIF to refine model reasoning, improve alignment with business expectations, and enable multi-step problem-solving across diverse tasks.

Hands-on experience building context-aware agents, integrating external knowledge sources through LlamaIndex, Neo4j, and enterprise data connectors, and developing rich user interactions through React-based agent interfaces.

Strong background in MLOps, DevOps, and scalable deployment, using Docker, Kubernetes, Azure DevOps, GitHub Actions, MLflow, DVC, and Terraform to manage CI/CD pipelines, experiment tracking, reproducibility, and infrastructure automation.

Built robust data ingestion and transformation pipelines utilizing Azure Databricks, PySpark, Azure Data Factory, AWS Glue, EMR, Lambda, and GCP DataStream — enabling near real-time processing, distributed compute, and high-performance ETL for AI workloads.

Developed full-stack machine learning systems using scikit-learn, XGBoost, LightGBM, TensorFlow, PyTorch, and Hugging Face Transformers, supporting use cases across NLP, classification, time-series forecasting, and pattern recognition.

Designed and deployed intelligent automation solutions that integrate AI agents with cloud platforms such as Azure ML, Databricks, AppSheet, and Power Automate, enabling automated triage, workflow routing, decisioning, and system troubleshooting.

Strong foundation in data visualization, analytics, and business intelligence, creating impactful dashboards with Power BI, Plotly, Matplotlib, Seaborn, and QuickSight, and using advanced statistical techniques (NumPy, Statsmodels, Prophet) for insights and forecasting.

Committed to responsible and explainable AI, applying SHAP, LIME, model fairness checks, drift monitoring, observability tools (CloudWatch, Azure Monitor, Grafana), and strong governance practices to ensure transparency, stability, and trustworthiness of ML and GenAI systems.

Technical Skills:

GenAI, LLMs & Multi-Agent Systems: LangGraph, LangChain, AutoGen, CrewAI, LlamaIndex, MCP, GPT-4 techniques, BERT, RoBERTa, Hugging Face Transformers, open-source LLMs, RLHF, RLAIF, prompt engineering, reasoning loops.

RAG, NLP & Vector Search: FAISS, Pinecone, Weaviate, ChromaDB, OpenSearch, embedding pipelines, document indexing, Neo4j, spaCy, tokenization methods, embedding strategies, transformer fine-tuning, entity extraction, document classification.

Machine Learning & Deep Learning: Scikit-learn, XGBoost, LightGBM, TensorFlow, PyTorch.

Data Engineering & Cloud Platforms: Azure Databricks, PySpark, Azure Data Factory, Synapse Analytics, Delta Lake, Spark SQL, SnowSQL, Snowpipe, AWS Glue, EMR, Athena, Lambda, SageMaker, GCP DataStream, GCP Cloud Logging.

MLOps, DevOps & Deployment: Docker, Kubernetes, MLflow, DVC, Terraform, SageMaker Pipelines, Azure DevOps, GitHub Actions, ECS.

Databases, Storage & Analytics: S3, Glue Catalog, Synapse SQL Pools, Azure SQL, Cosmos DB, Snowflake, Power BI, Plotly, Matplotlib, Seaborn, QuickSight, Pandas, Polars, NumPy, Statsmodels, Prophet.

Educational Details:

Master of Science (M.S.) in Management information system

University of Illinois Springfield, Springfield, IL USA Dec 2024

Bachelor of Technology (B.Tech.) in Electrical and Electronics Engineering

Swarna Bharathi Institute of Science & Technology, Khammam, India

Certifications:

Gen AI: Promt Engineering Basics - Coursera

Microsoft Power BI Developer (PL-300)

Microsoft AI & ML Engineering - Udemy

Professional Experience:

Smartly Feb 2024 – Current

Gen AI Developer Chicago, IL

Built multi-agent AI workflows by using LangGraph to create task flows between agents, applying AutoGen to implement reasoning loops, and using CrewAI to enable coordinated agent behavior with optional human oversight.

Developed context-aware agents by working with LangChain to structure prompts and logic, using LlamaIndex to connect agents to external knowledge sources, applying Neo4j for graph-based context, and using React to build interactive agent interfaces.

Customized model behavior by using MCP to extend model capabilities and applying GPT-4 techniques to create agents that can summarize content, analyze structured information, perform triage, and automate workflow steps.

Built memory-driven RAG pipelines by using FAISS for fast vector search, Pinecone for large-scale indexing, Weaviate for semantic retrieval, and ChromaDB for embedding storage to deliver low-latency document recall.

Improved agent reasoning by applying reinforcement learning techniques such as RLHF to align outputs with expected behavior and RLAIF to strengthen multi-step decision-making and adaptability.

Created modular agent flows using LangGraph, enabling agents to communicate, share context, discover new capabilities, and adapt to changing tasks in real time.

Developed troubleshooting and triage agents by using MCP adapters to access system tools, allowing the agents to run contextual queries, recommend fixes, and generate complete action steps with minimal human intervention.

Built workflow automation bots by using Azure Databricks to handle large-scale data operations, applying Azure Machine Learning to deploy AI components, and embedding them into AppSheet and Power Automate for automated routing and decisioning.

Developed ingestion and transformation pipelines by using Databricks for distributed data processing, applying PySpark to handle unstructured content, using Azure Data Factory for orchestration, and integrating GCP DataStream for near real-time data flow.

Implemented DevOps and MLOps pipelines by using Docker to containerize AI workloads, Kubernetes to run scalable deployments, Azure DevOps and GitHub Actions for CI/CD, and MLflow, DVC, and Terraform to manage experiments, versioning, and infrastructure.

Built monitoring solutions by using Azure Monitor to track system activity, applying Application Insights to observe agent performance and reasoning patterns, and using Cloud Logging to track latency, drift, and usage trends.

Contributed to GenAI platform development by building prototypes with LangGraph, using Neo4j for reasoning over relationships, creating interfaces with React, and deploying solutions using GCP services, helping transition them into stable, production-ready multi-agent systems.

HSBC Mar 2021 – Aug 2023

AI/ML Engineer Hyderabad, India

Built advanced LLM and early GenAI solutions by combining frameworks like LangChain for orchestration, Hugging Face Transformers for model loading and fine-tuning, and open-source models such as GPT. Used vector search engines like FAISS and OpenSearch to support high-accuracy semantic retrieval, summarization, and intelligent Q&A systems.

Designed Retrieval-Augmented Generation (RAG) architectures by pairing FAISS and OpenSearch vector indexes with LLMs, enabling models to pull in domain-aware context, reason over document embeddings, and deliver more accurate and grounded responses.

Developed end-to-end machine learning systems using tools such as Scikit-learn for classical modeling, XGBoost and LightGBM for boosted tree algorithms, and deep learning libraries like TensorFlow and PyTorch to train neural networks tailored to complex pattern-recognition tasks.

Created robust NLP pipelines by leveraging BERT and RoBERTa for transformer-based text understanding, using spaCy for linguistic preprocessing, and applying tokenization, embedding, and fine-tuning strategies to extract entities, classify documents, and interpret language at scale.

Built automated MLOps workflows using SageMaker Pipelines for orchestrating model life cycles, MLflow for experiment tracking and lineage, and Docker containers running on ECS for consistent and scalable model deployment across environments.

Engineered data ingestion and transformation pipelines with AWS Glue for ETL orchestration, EMR with PySpark for distributed compute, Athena for interactive querying, Lambda for serverless processing, and SageMaker for hosting data preparation jobs and integrating preprocessing into ML workflows.

Designed feature management systems using Glue Catalog for metadata governance, S3 for scalable storage, and custom registries to ensure that features remained versioned, reproducible, traceable, and ready for reuse across multiple ML models.

Implemented monitoring and observability layers by integrating CloudWatch metrics and logs, DynamoDB for lightweight state tracking, and Grafana dashboards to visualize model performance, latency, data drift, and health signals over time.

Performed deep statistical and analytical work using Pandas for data manipulation, Polars for high-performance processing, NumPy for numerical computing, Statsmodels for traditional statistical inference, and Prophet for time-series modeling—supporting trend discovery and predictive insight generation.

Built visual analytics through tools such as Plotly for interactive visualization, Matplotlib and Seaborn for exploratory graphics, and QuickSight for dynamic dashboards, enabling clear communication of insights to both technical and non-technical audiences.

Embedded responsible AI practices by applying SHAP and LIME to interpret model behavior, performing fairness and sensitivity checks, and ensuring every ML and GenAI component remained transparent, explainable, and suitable for governance reviews.

Ajanta Pharma May 2020 - Feb 2021

Data Analyst(Power BI) Hyderabad, India

Designed and delivered interactive Power BI dashboards featuring advanced DAX measures, drillthrough pages, custom visuals, and optimized data models to improve report performance and usability.

Built scalable data ingestion pipelines using Azure Data Factory (ADF), structuring scheduled loads, implementing control logic, and managing data transformations required for analytics.

Developed complex transformation logic in Azure Databricks using PySpark and Spark SQL, cleansing and enriching raw data to create reliable, analysis-ready datasets.

Leveraged Azure Synapse Analytics for additional orchestration and SQL transformations, supporting large-scale workloads and performance-driven data modeling.

Designed structured analytical datasets using star schema modeling and Delta Lake storage patterns to provide consistent and reusable data models for reporting.

Integrated data from multiple environments, including Azure SQL, Synapse SQL Pools, Snowflake, Cosmos DB, AWS S3, and on-premises systems, ensuring robust and smooth data ingestion.

Published and maintained content in the Power BI Service, organizing workspaces, configuring dataflows, establishing refresh pipelines, and supporting governed access to shared datasets.

Automated ETL and ELT processes using Databricks notebooks, AWS Glue jobs, and SnowSQL/Snowpipe scripts, improving data reliability and minimizing manual interventions.

Created monitoring dashboards in Power BI, integrating data from Azure Log Analytics to track pipeline health, data freshness, performance bottlenecks, and system stability.

Collaborated closely with business and technical teams to gather requirements, interpret analytical needs, and translate them into meaningful data models, dashboards, and insights.

Supported CI/CD processes using Azure DevOps, building deployment pipelines for ADF, Databricks, Synapse, DBT, and Power BI assets, and writing PowerShell scripts for automation and documentation.

Contact this candidate