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

I am a results-driven senior AI/ML engineer with 9 years of cross-functional experience spanning MLOps, LLMOps, deep learning, classical ML, and AI platform engineering. I have led full lifecycle ML solutions from model development to scalable deployment across healthcare, financial services, retail, and enterprise data environments. I operationalize GenAI and ML systems on GCP, Azure, and AWS using robust CI/CD, Kubernetes orchestration, and containerized microservices. I focus on delivering reproducible, explainable, and secure AI systems with a strong foundation in Python, model serving, infrastructure as code, and regulatory compliance. My track record includes building modular RAG pipelines, enabling multi-LLM routing and vector indexing; deploying GenAI inference services with BentoML, KServe, and GKE; implementing observability with MLflow, DVC, PromptLayer, and Evidently AI; and orchestrating end-to-end LLMOps patterns for reliability, governance, and auditability across production workflows.…I am a results-driven senior AI/ML engineer with 9 years of cross-functional experience spanning MLOps, LLMOps, deep learning, classical ML, and AI platform engineering. I have led full lifecycle ML solutions from model development to scalable deployment across healthcare, financial services, retail, and enterprise data environments. I operationalize GenAI and ML systems on GCP, Azure, and AWS using robust CI/CD, Kubernetes orchestration, and containerized microservices. I focus on delivering reproducible, explainable, and secure AI systems with a strong foundation in Python, model serving, infrastructure as code, and regulatory compliance. My track record includes building modular RAG pipelines, enabling multi-LLM routing and vector indexing; deploying GenAI inference services with BentoML, KServe, and GKE; implementing observability with MLflow, DVC, PromptLayer, and Evidently AI; and orchestrating end-to-end LLMOps patterns for reliability, governance, and auditability across production workflows.

Deepthi Chava

AI Engineer, Data Scientist, Full Stack Developer, +2





I am a results-driven senior AI/ML engineer with 9 years of cross-functional experience spanning MLOps, LLMOps, deep learning, classical ML, and AI platform engineering. I have led full lifecycle ML solutions from model development to scalable deployment across healthcare, financial services, retail, and enterprise data environments. I operationalize GenAI and ML systems on GCP, Azure, and AWS using robust CI/CD, Kubernetes orchestration, and containerized microservices. I focus on delivering reproducible, explainable, and secure AI systems with a strong foundation in Python, model serving, infrastructure as code, and regulatory compliance. My track record includes building modular RAG pipelines, enabling multi-LLM routing and vector indexing; deploying GenAI inference services with BentoML, KServe, and GKE; implementing observability with MLflow, DVC, PromptLayer, and Evidently AI; and orchestrating end-to-end LLMOps patterns for reliability, governance, and auditability across production workflows.…I am a results-driven senior AI/ML engineer with 9 years of cross-functional experience spanning MLOps, LLMOps, deep learning, classical ML, and AI platform engineering. I have led full lifecycle ML solutions from model development to scalable deployment across healthcare, financial services, retail, and enterprise data environments. I operationalize GenAI and ML systems on GCP, Azure, and AWS using robust CI/CD, Kubernetes orchestration, and containerized microservices. I focus on delivering reproducible, explainable, and secure AI systems with a strong foundation in Python, model serving, infrastructure as code, and regulatory compliance. My track record includes building modular RAG pipelines, enabling multi-LLM routing and vector indexing; deploying GenAI inference services with BentoML, KServe, and GKE; implementing observability with MLflow, DVC, PromptLayer, and Evidently AI; and orchestrating end-to-end LLMOps patterns for reliability, governance, and auditability across production workflows.

Available to hire

I am a results-driven senior AI/ML engineer with 9 years of cross-functional experience spanning MLOps, LLMOps, deep learning, classical ML, and AI platform engineering. I have led full lifecycle ML solutions from model development to scalable deployment across healthcare, financial services, retail, and enterprise data environments. I operationalize GenAI and ML systems on GCP, Azure, and AWS using robust CI/CD, Kubernetes orchestration, and containerized microservices.

I focus on delivering reproducible, explainable, and secure AI systems with a strong foundation in Python, model serving, infrastructure as code, and regulatory compliance. My track record includes building modular RAG pipelines, enabling multi-LLM routing and vector indexing; deploying GenAI inference services with BentoML, KServe, and GKE; implementing observability with MLflow, DVC, PromptLayer, and Evidently AI; and orchestrating end-to-end LLMOps patterns for reliability, governance, and auditability across production workflows.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Work Experience

Lead AI/ML Engineer – Gen AI at Cotiviti

February 1, 2025 - Present

Architected multi-LLM Retrieval-Augmented Generation pipelines using LangChain, FAISS, and Vertex AI to enable context-aware audit automation for payer datasets with domain-tuned LLaMA2 and PaLM models. Containerized and deployed LLM microservices with BentoML and KServe on GKE for high availability and GPU-efficient production endpoints. Designed a vector-store abstraction layer (Pinecone, Qdrant, Weaviate) for modular swapping of similarity search across claim summarization and patient explanations. Fine-tuned Mistral-7B and LLaMA2 with Hugging Face Transformers and tracked lineage with MLflow and DVC for reproducibility. Integrated PromptLayer with LangChain to monitor prompt quality and token usage across agents supporting billing code analysis. Built a hybrid FAISS/Elasticsearch search stack to blend dense retrieval with metadata filtering over large healthcare corpora. Deployed CLIP and Whisper for multimodal workflows (invoicing, OCR explanations, document review). Orchestrated

Senior ML Engineer at Mantech

January 1, 2025 - October 15, 2025

Architected and scaled secure, explainable ML deployment infrastructures for mission-critical cybersecurity analytics on Azure. Delivered CI/CD-enabled ML workflows, feature stores, and inference services with operational visibility and regulatory compliance. Built reproducible pipelines with MLflow, DVC, and Apache Airflow; containerized models with TorchServe, TensorFlow Serving, and Triton Inference Server; orchestrated AKS-based deployments with Helm, enabling autoscaling and secure exposure. Provisioned IaC with Terraform to automate AKS clusters, Azure Key Vault secrets, storage accounts, and policies. Implemented real-time and batch feature pipelines with Feast and Azure storage; automated CI/CD for model training, validation, and deployment using GitHub Actions with gated promotions. Integrated Azure ML SDK for orchestrating training, managing compute, model registration, and endpoint exposure. Added SHAP and LIME explainability, Argo-compatible deployment manifests, and end-to

ML Platform Engineer at USAA

September 1, 2023 - October 15, 2025

Built reproducible ML pipelines with DVC and MLflow to enable dataset, model, and metric versioning for financial risk and fraud analytics. Containerized models with Docker and deployed via Kubernetes, exposing scalable microservices with pre/post inference hooks. Implemented CI/CD with GitHub Actions and Jenkins, enforcing gated promotions and environment isolation. Managed model lifecycle with MLflow Tracking and MLflow Registry, integrating it with GitOps for stage-wise approvals and rollback. Configured SageMaker training/hosting with REST APIs for secure fraud scoring. Integrated DVC for feature data provenance and lineage, and provisioned Kubernetes resources with Helm for isolated experimentation. Implemented automated drift detection hooks in CI pipelines, and refactored notebooks into production-grade modules coordinated through Jupyter/CI. Collaborated with platform security to enforce container scanning, image whitelisting, and IAM controls. Standardized Python environments

Machine Learning Engineer at Vertex

March 1, 2021 - October 15, 2025

Delivered ML-driven customer intelligence pipelines focusing on churn risk scoring, segmentation, and campaign optimization. Built supervised models (Logistic Regression, Random Forest, XGBoost) in scikit-learn; engineered data preprocessing with Pandas/NumPy and feature engineering (time-based features, encoding). Developed unsupervised segmentation with K-Means and PCA; trained deep neural networks with Keras on high-cardinality features. Conducted hyperparameter tuning via GridSearchCV with cross-validation, optimizing ROC-AUC, F1, and LogLoss; applied SMOTE to address class imbalance; designed evaluation workflows with ROC-AUC and PR curves; converted semi-structured logs to structured data, enabling automated feature extraction and schema enforcement. Orchestrated batch inference with Python CLIs and shell scripts; maintained notebooks as production-grade modules and ensured reproducible experiments.

Python Developer at Object Computing

October 1, 2019 - October 15, 2025

Developed and maintained ETL pipelines on Oracle and SQL Server to support enterprise applications. Automated data ingestion, normalization, and validation with Python (Pandas, NumPy). Designed parameterized SSRS reports and SQL views for auditable multi-unit reporting. Built reusable data models and SQL scripts for standardized analytics; optimized heavy join queries and materialized views to reduce batch runtimes. Performed data profiling, anomaly detection, and quality checks; documented ETL processes, mappings, and data lineage. Standardized environments and tooling to ensure consistent, maintainable pipelines across client projects.