Skip to content

Tarun Kumar Reddy Nallagari

AI Engineer • New York, NY, USA • n**************@gmail.com • 716****536 • linkedin.com/••••• • drivetube.ai/•••••

Professional Summary

AI Engineer with 4+ years of experience building production machine learning and Generative AI systems across financial services and healthcare. Hands-on with LLMs, Retrieval-Augmented Generation, vector databases, prompt engineering, end-to-end ML pipelines, MLOps, and real-time inference on AWS. AWS Certified Data Engineer with experience operationalizing models in regulated environments and driving audit-ready model governance.

Technical Skills

Programming Languages: Python
Web Technologies: REST APIs
Frameworks and Libraries: LangChain,scikit-learn,PyTorch
Databases: SQL,Snowflake
Cloud and DevOps: Docker,Kubernetes,MLflow,Apache Airflow,AWS SageMaker,AWS Glue,Azure Databricks
Data and Analytics: Feature Engineering,Predictive Modeling,XGBoost
Skills: PySpark
Generative AI & LLM Systems: LLMs,Retrieval-Augmented Generation,Pinecone,FAISS,Embeddings,Prompt Engineering

Work Experience

Ally Financial
New York, USA
AI Engineer
Jul 2025 – Present
Worked at Ally Financial (banking and auto-finance) delivering production ML and Generative AI solutions for credit risk, fraud detection, document processing, and customer support automation.
Tech Stack: Python, LangChain, Pinecone, FAISS, AWS SageMaker, AWS Glue, Docker, Kubernetes, MLflow, Apache Airflow, REST APIs, scikit-learn, XGBoost, PyTorch
  • Built Retrieval-Augmented Generation (RAG) pipelines with LangChain and vector DBs (Pinecone, FAISS) to ground LLM responses in enterprise banking documents, improving grounded-response accuracy by 30% for document-processing and customer-support use cases.
  • Benchmarked LLM APIs, embedding models, and vector stores and implemented prompt-engineering plus evaluation workflows, reducing hallucinated outputs by 25% in sampled production reviews and guiding vendor/model selection.
  • Designed and deployed Python-based production ML models for credit-risk scoring and fraud detection across auto-finance and banking products, increasing detection precision by 20% versus prior rule-based systems.
  • Built end-to-end ML pipelines and CI/CD automation using SageMaker, MLflow, and Airflow to manage feature engineering, training, evaluation, and deployment, cutting model release-cycle time by 30%.
  • Containerized and served ML models with Docker and Kubernetes on AWS, exposing REST APIs to downstream systems and reducing p95 prediction latency by 25% for low-latency real-time decisioning.
  • Implemented model monitoring and drift-detection workflows and automated retraining, aligning with SR 11-7 model governance practices to improve audit readiness and reduce production performance incidents by 25%.
Optum
Bengaluru, India
Data Engineer
Mar 2021 – Dec 2023
Worked at Optum (healthcare services) building data engineering pipelines, NLP feature extraction, and analytics platforms for claims, clinical data, and population health programs.
Tech Stack: Python, PySpark, Apache Spark, Azure Databricks, Snowflake, SQL, Apache Airflow, Hadoop
  • Applied NLP and information-extraction techniques to convert unstructured clinical notes and provider documentation into structured features, improving data usability for risk-adjustment and HEDIS workflows by 25%.
  • Developed ML models to predict member health-risk scores and surface care gaps, supporting care-management and population-health programs and improving risk-stratification accuracy for downstream care teams.
  • Built and maintained large-scale ETL pipelines in Python and PySpark for healthcare claims, member eligibility, and clinical datasets, increasing daily data throughput by 30% for analytics and data science consumers.
  • Designed and optimized SQL queries and data models over multi-terabyte claims warehouses on Azure Databricks and Snowflake, reducing reporting and query latency by 25% for downstream reporting teams.
  • Implemented automated data-validation and data-quality frameworks enforcing HIPAA-compliant healthcare data standards, reducing downstream data errors by 20% across analytics pipelines.
  • Orchestrated recurring jobs using Apache Airflow and containerized workloads, improving pipeline reliability and reducing manual intervention while partnering with data scientists to productionize features and models.

Education

State University of New York at Buffalo
Master of Science in Data Science • Buffalo, NY • Jan 2024 – Jun 2025

Certifications

AWS Certified Data Engineer – Associate — Amazon Web Services (AWS)

Powered by Drivetube · Create your own profile at drivetube.ai