Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I'm Karthik Nagula, a Senior AI/ML Engineer with 10+ years of experience designing, building, and deploying advanced LLM and multimodal AI systems across healthcare, banking, telecom, and enterprise domains, with a strong focus on production-grade scalability and real-world impact. I have hands-on experience delivering end-to-end LLM and Vision-Language Model pipelines, robust evaluation and monitoring frameworks, and real-time inference systems, enabling measurable business outcomes such as $187M in healthcare savings and $285M in fraud loss prevention. I thrive on building multi-agent AI workflows with LangGraph, integrating with enterprise data sources, and delivering reliable, low-latency AI services in production.…I'm Karthik Nagula, a Senior AI/ML Engineer with 10+ years of experience designing, building, and deploying advanced LLM and multimodal AI systems across healthcare, banking, telecom, and enterprise domains, with a strong focus on production-grade scalability and real-world impact. I have hands-on experience delivering end-to-end LLM and Vision-Language Model pipelines, robust evaluation and monitoring frameworks, and real-time inference systems, enabling measurable business outcomes such as $187M in healthcare savings and $285M in fraud loss prevention. I thrive on building multi-agent AI workflows with LangGraph, integrating with enterprise data sources, and delivering reliable, low-latency AI services in production.

Karthik Nagula

Data Scientist, AI Engineer, Full Stack Developer, +4





I'm Karthik Nagula, a Senior AI/ML Engineer with 10+ years of experience designing, building, and deploying advanced LLM and multimodal AI systems across healthcare, banking, telecom, and enterprise domains, with a strong focus on production-grade scalability and real-world impact. I have hands-on experience delivering end-to-end LLM and Vision-Language Model pipelines, robust evaluation and monitoring frameworks, and real-time inference systems, enabling measurable business outcomes such as $187M in healthcare savings and $285M in fraud loss prevention. I thrive on building multi-agent AI workflows with LangGraph, integrating with enterprise data sources, and delivering reliable, low-latency AI services in production.…I'm Karthik Nagula, a Senior AI/ML Engineer with 10+ years of experience designing, building, and deploying advanced LLM and multimodal AI systems across healthcare, banking, telecom, and enterprise domains, with a strong focus on production-grade scalability and real-world impact. I have hands-on experience delivering end-to-end LLM and Vision-Language Model pipelines, robust evaluation and monitoring frameworks, and real-time inference systems, enabling measurable business outcomes such as $187M in healthcare savings and $285M in fraud loss prevention. I thrive on building multi-agent AI workflows with LangGraph, integrating with enterprise data sources, and delivering reliable, low-latency AI services in production.

Available to hire

I’m Karthik Nagula, a Senior AI/ML Engineer with 10+ years of experience designing, building, and deploying advanced LLM and multimodal AI systems across healthcare, banking, telecom, and enterprise domains, with a strong focus on production-grade scalability and real-world impact.

I have hands-on experience delivering end-to-end LLM and Vision-Language Model pipelines, robust evaluation and monitoring frameworks, and real-time inference systems, enabling measurable business outcomes such as $187M in healthcare savings and $285M in fraud loss prevention. I thrive on building multi-agent AI workflows with LangGraph, integrating with enterprise data sources, and delivering reliable, low-latency AI services in production.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

Senior Machine Learning & Conversational AI Engineer at Humana

September 1, 2024 - Present

Architected and deployed cloud-native conversational AI platforms on Google Cloud using Vertex AI to support intelligent virtual agents for healthcare engagement, automation, and decision-support workflows. Implemented advanced conversational flows with intent detection, context management, and slot filling; built retrieval augmented generation pipelines grounding responses in clinical docs and enterprise data; established observability and MLOps processes; deployed on Kubernetes with secure REST APIs.

Senior Machine Learning & NLP Engineer (Conversational AI) at U.S. Bank

November 1, 2022 - August 1, 2024

Designed and implemented conversational AI solutions on Google Cloud using Vertex AI and Dialogflow to support customer interaction and internal investigation workflows in regulated financial environments. Built intent classification and entity extraction models; integrated LLM-based reasoning and summarization; developed Retrieval Augmented Generation pipelines grounding responses in financial data; automated model training and deployment with Vertex AI Pipelines.

Machine Learning & Conversational AI Engineer at Cisco

May 1, 2020 - October 1, 2022

Designed and developed cloud-native conversational AI services on Google Cloud using Vertex AI to support enterprise knowledge discovery and automated decision-support workflows. Built Retrieval Augmented Generation pipelines; orchestrated tool-driven agent flows; leveraged transformer-based NLP models for improved intent classification and contextual reasoning; deployed on GKE with secure REST APIs and observability.

Python & Conversational AI Engineer at AT&T

September 1, 2018 - April 1, 2020

Developed conversational AI solutions to automate telecom incident analysis and customer interaction workflows using NLP, intent classification, and context-aware dialogue management. Implemented retrieval-based flows with semantic similarity, built Python-based fulfillment services, and deployed containerized agents with monitoring and alerting for high-volume interactions.

Python & Machine Learning Engineer at GAP Inc

June 1, 2015 - July 1, 2018

Developed ML-driven backend services supporting intelligent retail analytics, recommendations, and data-driven decision workflows. Built NLP pipelines for semantic search across catalogs, engineered data preprocessing and feature pipelines, and deployed scalable ML services with monitoring for inference performance and data quality.

Senior AI/ML engineer at HUMANA

September 1, 2024 - Present

Led delivery of the AI/ML claims intelligence platform across v1–v3 in an Agile SDLC, boosting experiment-to-production turnaround by ~31%. Architected Retrieval-Augmented Generation (RAG) solutions grounded in medical policies and clinical notes, improving first-pass adjudication accuracy by ~24%. Integrated GPT-based LLMs via Azure OpenAI to generate clinically grounded summaries, reducing manual reviewer effort by ~38% and increasing claims throughput by ~32%. Implemented deterministic multi-step LLM orchestration with LangChain/LangGraph, tuned vector retrieval with FAISS and embeddings, and built robust data pipelines for ICD-10, CPT, utilization, and provider behavior data. Deployed REST inferences with FastAPI, enforced prompt governance, and established full model lifecycle management with containerized artifacts and automated retraining. Enhanced observability and explainability with Datadog/Prometheus and SHAP-style attribution, reducing production incidents by ~28% and imp

AI / ML Engineer at U.S. Bank

November 1, 2022 - August 1, 2024

Delivered a fraud detection platform at scale for ~2.1B monthly transactions. Built a data foundation on Snowflake and engineered ~18B feature vectors monthly with Apache Spark, enabling rapid experimentation and consistent batch/real-time operation. Implemented real-time scoring with Kafka ingestion (~70K TPS) and low-latency inference (<100 ms p95) via Triton and FastAPI. Achieved high fraud detection performance (AUC 0.96 for transaction-level models; 0.94 precision for behavioral/account takeover models) and advanced NLP signals with BERT Sentence Transformers. Managed end-to-end model lifecycle with SageMaker/Bedrock, MLflow, and Databricks, plus extensive monitoring with Evidently AI and Splunk for auditability and SOX compliance.

Data science & ML engineer at CISCO

May 1, 2020 - October 1, 2022

Delivered multi-release network telemetry ML capabilities, improving anomaly detection precision by ~28% and reducing manual triage by ~35%. Built scalable feature pipelines for time-series KPIs and event streams, enabling stable training and deployment via Vertex AI Pipelines. Implemented real-time prediction services with FastAPI/Docker/Kubernetes, standardized CI/CD, and added explainability outputs. Enhanced security and governance with RBAC and centralized logging. Optimized data processing and compute sizing to improve latency and reduce costs while maintaining reliability.

Data Scientist at AT&T

September 1, 2018 - April 1, 2020

Built an AI Assistant Platform (v1–v3) powering conversational analytics and guided insights across customer usage and network performance. Integrated GPT-based LLMs via AWS and LangChain to support natural language interactions, explanations, and summaries, increasing accessibility of telecom insights. Implemented multi-step prompt tuning (Chain-of-Thought) and structured workflows for tool invocation and response validation. Deployed containerized services on AWS with monitoring and observability to ensure reliability during peak usage.

Data engineer at GAP Inc

June 1, 2015 - July 1, 2018

Delivered an AI Automation Platform (v1–v3) for ingesting, validating, transforming, and publishing retail data, improving data availability by ~35% and reducing manual handling by ~30%. Implemented metadata interpretation with early LLMs, versioned REST APIs, and RBAC-based security. Established data pipelines in Python, with data warehousing on Azure and robust runbooks for onboarding, change management, and auditability.

AI/ML Engineer at U.S. Bank

November 1, 2022 - August 1, 2024

Delivered a real-time fraud detection platform using Agile SDLC, supporting incremental model releases across a highly regulated banking ecosystem processing ~2.1B transactions monthly while maintaining alignment with PCI-DSS and SOX compliance frameworks. Centralized fraud data platform on Snowflake Enterprise Edition, ingesting ~5TB of transactional and behavioral data daily. Implemented real-time ingestion pipelines using Apache Kafka (70K TPS) and developed fraud detection models with XGBoost/LightGBM (AUC 0.96). Enhanced fraud intelligence with BERT Sentence Transformers for unstructured narratives and designed hybrid retrieval using BM25 + pgvector, improving top-k relevance and reducing hallucinations. Operationalized model training, validation, and deployment on AWS Bedrock and SageMaker, with observability via CloudWatch, X-Ray, and Splunk.