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

I'm Ovian Thomas, a Senior AI/ML Engineer specializing in production-grade LLM systems, RAG, and agent workflows. I translate ambiguous AI capabilities into dependable, scalable solutions for healthcare, enterprise software, and cloud platforms. I collaborate across teams to design end-to-end AI workflows, implement robust observability, optimize retrieval and latency, and ship features that work in the real world. I enjoy turning complex requirements into maintainable architectures and fostering reliable, governance-conscious AI systems.…I'm Ovian Thomas, a Senior AI/ML Engineer specializing in production-grade LLM systems, RAG, and agent workflows. I translate ambiguous AI capabilities into dependable, scalable solutions for healthcare, enterprise software, and cloud platforms. I collaborate across teams to design end-to-end AI workflows, implement robust observability, optimize retrieval and latency, and ship features that work in the real world. I enjoy turning complex requirements into maintainable architectures and fostering reliable, governance-conscious AI systems.

Ovian Thomas

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





I'm Ovian Thomas, a Senior AI/ML Engineer specializing in production-grade LLM systems, RAG, and agent workflows. I translate ambiguous AI capabilities into dependable, scalable solutions for healthcare, enterprise software, and cloud platforms. I collaborate across teams to design end-to-end AI workflows, implement robust observability, optimize retrieval and latency, and ship features that work in the real world. I enjoy turning complex requirements into maintainable architectures and fostering reliable, governance-conscious AI systems.…I'm Ovian Thomas, a Senior AI/ML Engineer specializing in production-grade LLM systems, RAG, and agent workflows. I translate ambiguous AI capabilities into dependable, scalable solutions for healthcare, enterprise software, and cloud platforms. I collaborate across teams to design end-to-end AI workflows, implement robust observability, optimize retrieval and latency, and ship features that work in the real world. I enjoy turning complex requirements into maintainable architectures and fostering reliable, governance-conscious AI systems.

Available to hire

I’m Ovian Thomas, a Senior AI/ML Engineer specializing in production-grade LLM systems, RAG, and agent workflows. I translate ambiguous AI capabilities into dependable, scalable solutions for healthcare, enterprise software, and cloud platforms.

I collaborate across teams to design end-to-end AI workflows, implement robust observability, optimize retrieval and latency, and ship features that work in the real world. I enjoy turning complex requirements into maintainable architectures and fostering reliable, governance-conscious AI systems.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

AI Systems Architect at Anthropic

June 1, 2025 - February 1, 2026

Designed and improved production LLM systems for enterprise-use cases, focusing on agent workflows, retrieval quality, tool safety, and inference observability across cloud environments. Built multi-step Claude-based workflows for document and API-driven tasks using structured tool invocation and graph-backed context, improving reliability across internal enterprise workflows. Helped define MCP-based integration patterns for enterprise data access, enabling internal tools and services to expose context to LLM applications more consistently. Developed multimodal document-analysis and retrieval pipelines combining OCR, text chunking, and image-aware context for complex PDFs and visually structured documents. Added guardrails around tool execution and permissioned actions, including policy checks, approval flows, and safer handling of high-impact operations. Improved observability for AWS- and GCP-hosted inference workloads through tracing, latency monitoring, and prompt-cache instrumenta

Senior AI/ML Engineer at Anthropic

January 1, 2023 - May 1, 2025

Focused on RAG, tool-using agent workflows, and production AI systems, owning orchestration, evaluation, and reliability standards for customer-facing experiences and enterprise integrations. Engineered Contextual Retrieval (RAG) systems using Python, Pinecone, LangChain, and LlamaIndex, with metadata filtering and chunk-specific context to improve retrieval precision and answer verifiability for Claude’s high-context window. Developed tool-use interfaces using JSON Schema and structured tool contracts, improving reliability for function-calling and API-driven workflows. Built orchestration workflows with LangGraph-style state management, including task planning, retry logic, and tool-result validation, improving completion rates by 25% on multi-step internal and customer-facing tasks. Evaluated and tested multi-agent workflows using frameworks such as LangGraph and CrewAI, ensuring robust task coordination and performance across Claude 3.5-based agent systems. Built integrations and

Senior Machine Learning Engineer at Amazon Web Services (AWS)

April 1, 2020 - December 1, 2022

Contributed to multimodal computer vision systems for Just Walk Out, supporting real-time multi-object tracking and perception across multi-camera retail environments. Supported high-concurrency inference workflows for computer vision systems, improving robustness and latency in production retail deployments. Built and maintained end-to-end ML pipelines using Amazon SageMaker Pipelines and Model Registry for training, validation, deployment, and versioning of models. Developed document and image processing workflows using Amazon Textract and Amazon Rekognition, enabling scalable extraction, classification, and analysis of structured and unstructured data. Designed and deployed containerized inference services using Docker, Kubernetes (Amazon EKS), and REST/gRPC APIs, improving deployment consistency and scalability. Implemented MLOps best practices including CI/CD (GitHub Actions / AWS CodePipeline), monitoring, logging, and model validation, improving production reliability and releas

Machine Learning Engineer at Flatiron Health

September 1, 2018 - April 1, 2020

Developed predictive models on structured and unstructured oncology EHR data using Python, scikit-learn, and feature engineering, improving early risk detection accuracy by about 15% on defined patient cohorts. Built NLP pipelines for clinical note extraction using tokenization, normalization, and rule-based methods, increasing feature coverage and improving downstream model performance. Designed and maintained ETL pipelines (Python, SQL) with HIPAA-compliant access controls, logging, and data validation, ensuring reliable processing of large-scale clinical datasets. Implemented model interpretability workflows using SHAP and feature attribution, improving transparency and enabling clinician-facing review of model outputs. Reduced data preparation time by about 20% through pipeline optimization, reusable feature transformations, and automated validation checks, accelerating experimentation and model iteration. Collaborated with clinicians, analysts, and product teams to translate oncol

Software Engineer at Accenture

August 1, 2016 - August 1, 2018

Built and maintained REST APIs and backend services using Java, Spring Boot, and SQL, supporting enterprise applications with strict integration and reliability requirements. Refactored legacy systems into modular, service-oriented architectures (SOA) with versioned API contracts, improving maintainability and reducing deployment risk. Supported migration of on-premise applications to AWS, including environment setup, deployment workflows, and validation processes, improving scalability and system consistency. Developed data processing and reporting pipelines using SQL and scheduled backend jobs, improving data availability and operational visibility for business teams. Implemented CI/CD pipelines using Jenkins, introducing automated build, test, and deployment workflows that reduced release errors and improved delivery reliability. Collaborated with cross-functional teams and client stakeholders to define API contracts, migration plans, and operational runbooks, ensuring smooth execut