As an AI/ML Engineer, I specialize in designing and deploying enterprise-level Large Language Model (LLM) pipelines and Retrieval-Augmented Generation (RAG) systems.
Experience
2025 — Now
2025 — Now
• Designed, fine-tuned, and deployed enterprise-grade Large Language Model (LLM) pipelines using the OpenAI API, LangChain, and Python, leveraging advanced prompt engineering to automate customer behavior analysis, summarization, and product recommendation workflows, cutting manual processing time by 40%.
• Built and integrated Retrieval-Augmented Generation (RAG) systems leveraging vector databases (Pinecone, FAISS) to improve contextual accuracy, reasoning, and recall in AI responses, boosting precision by 30%.
• Developed scalable AI microservices using FastAPI, Docker, and AWS (Lambda, SageMaker, ECS), enabling secure, modular, and production-ready GenAI deployments across multiple domains.
• Implemented agentic AI workflows using LangGraph, AutoGen, and OpenAI function calling for autonomous task execution, memory management, and multi-agent orchestration, improving task completion efficiency by 35%.
• Established LLMOps monitoring frameworks using Weights & Biases, MLflow, and Prometheus, tracking drift, accuracy, latency, and cost optimization, reducing production issues by 25% and enhancing response speed by 40%.
• Collaborated cross-functionally with data engineering and backend teams to implement data ingestion, preprocessing, and vector embedding pipelines for structured and unstructured data sources.
2024 — 2024
2024 — 2024
• Architected a LangChain-based Retrieval-Augmented Generation (RAG) pipeline leveraging OpenAI GPT-4, FAISS, and ChromaDB for semantic search across 2M+ sales and financial records, reducing query latency by 70% and improving contextual accuracy by 60%.
• Developed a ReAct-driven AI agent for natural-language-to-SQL translation using GPT embeddings, automating 50+ complex analytics queries and accelerating decision-making cycles across departments.
• Deployed AI services using FastAPI, exposing REST endpoints with auto-generated Swagger UI (Open API) documentation to ensure integration readiness and maintainability.
• Unified RAG insights, GPT-4 outputs, and SageMaker analytics within interactive Power BI dashboards, empowering stakeholders with real-time, data-driven decision intelligence.
• Rapidly prototyped and deployed a Dockerized FastAPI and Streamlit interface for real-time SQL execution and KPI analytics, increasing adoption among analysts and executives by 45%.
• Integrated guardrails and schema validation to enforce structured outputs, prevent SQL injection risks, and improve system safety.
• Automated large-scale ETL pipelines using Python, PySpark, and AWS SageMaker Processing Jobs to ingest Salesforce API and IoT data into Amazon S3, achieving 99% data quality and 35% faster throughput.
• Translated complex Power BI DAX queries into Python scripts and scheduled them as automated jobs, generating recurring financial and operational reports without manual intervention.
2020 — 2022
2020 — 2022
• Designed, developed, and deployed predictive ML models (Regression, Classification, Clustering, XGBoost) to forecast customer churn and campaign performance, boosting retention by 22% and optimizing marketing ROI.
• Engineered and maintained end-to-end data pipelines leveraging SQL, Pandas, NumPy, Spark, and Airflow, enabling efficient processing of large-scale enterprise datasets and reducing data latency by 30%.
• Conducted in-depth exploratory data analysis (EDA), hypothesis testing (A/B testing, chi-square, ANOVA), and feature engineering to identify key behavioral drivers of churn and shape strategic interventions.
• Applied model evaluation, explainability, and validation techniques (cross-validation, SHAP/LIME) to ensure model robustness, interpretability, and compliance standards.
• Applied anomaly detection algorithms (Isolation Forest, DBSCAN) to uncover unusual customer behavior and segmentation anomalies, improving risk flagging accuracy and model robustness.
• Collaborated with cross-functional teams (product, marketing, data engineering, and business stakeholders) to translate requirements into ML solutions, driving automation, cost savings, and improved decision-making.
• Delivered client-facing dashboards using PowerBI and insights using Pandas, NumPy, Matplotlib, and Seaborn, driving data-backed strategic decisions.
2018 — 2020
2018 — 2020
• Extracted, processed, and validated 500K+ activation records from FTP sources into SQL Server pipelines, ensuring 99% data availability for downstream systems and reporting.
• Orchestrated and monitored 50+ ETL workflows using CA Autosys, implementing job scheduling, retries, and performance tuning to achieve 99.9% ingestion reliability.
• Developed and optimized ETL pipelines and SQL/NoSQL queries, enabling faster data access and reducing query latency by 25% across MySQL, PostgreSQL, and MongoDB.
• Streamlined cross-system ingestion workflows using SSIS, improving end-to-end data processing stability and reducing recurring pipeline failures.
• Designed and implemented SQL stored procedures for device return eligibility and exception handling, reducing manual workload by 40% and improving operational turnaround time.
• Optimized SQL queries and scheduling logic, cutting execution time by up to 30% and accelerating analytics delivery to business stakeholders.
• Automated data validation and monitoring using Python, SQL logging/alerts (triggers, Database Mail) and built Tableau dashboards, reducing reporting discrepancies by 25% and improving issue resolution time by 50%.
Education
University of Maryland Baltimore County
Master of Science - MS
Lakireddy Bali Reddy College of Engineering(Autonomous)