Senior Software Engineer AI & MLOps

Robert Romano

Senior Software Engineer

******.**.***@*****.*** 650-***-**** Newark, NJ www.linkedin.com/in/robert-romano-ab504420/ SUMMARY

Senior Software Engineer with 10+ years of experience specializing in scalable backend systems and the design, development, and deployment of enterprise AI and machine learning platforms. Proven track record building high-volume AI workflows, NLP-driven services, and computer vision pipelines using Python, TensorFlow, PyTorch, and cloud platforms (AWS, Azure), with a strong focus on performance, reliability, and integration. Expertise in full-stack AI development, from model training and optimization to production deployment on Kubernetes, event-driven microservices, and robust MLOps practices for business-critical systems. SKILLS

- AI & Machine Learning Development: Python, TensorFlow, PyTorch, Keras, Natural Language Processing (NLP), Computer Vision, Machine Learning Algorithms, Statistical Modeling, Model Optimization, Model Deployment

- Cloud & AI Platforms: AWS SageMaker, Azure OpenAI, Google Cloud AI Platform, AWS Lambda, Azure Functions, Containerized AI Serving

- Engineering & Infrastructure: Distributed Systems, Microservices Architecture, Event-Driven Systems, API Design, System Scalability, Fault Tolerance

- Programming Languages & Frameworks: C#, .NET, Python, FastAPI, Flask, ASP.NET Core

- Cloud & Orchestration: Microsoft Azure, AWS, Azure Kubernetes Service (AKS), Amazon EKS, Docker, Kubernetes, Terraform, Helm

- Data & ML Systems: SQL Server, PostgreSQL, Redis, Vector Search, Document Indexing Pipelines, Model Orchestration, Retrieval Augmented Generation (RAG)

- DevOps & MLOps: CI/CD Pipelines, GitHub Actions, Azure DevOps, Containerized Deployments, Release Automation, ML Model Lifecycle Management

- Monitoring & Observability: Datadog, Prometheus, OpenTelemetry, Metrics Pipelines, Model Performance Monitoring

- Domain Expertise: Enterprise SaaS, Cloud Platforms, AI Systems, Healthcare Systems, Observability Platforms PROFESSIONAL EXPERIENCE

Datadog Remote

Senior Software Engineer Jul 2023 – Present

- Architected and developed AI/ML-powered observability services using Python and TensorFlow to process high- volume infrastructure telemetry streams, reducing alert noise by 40% through advanced anomaly detection models.

- Engineered a real-time computer vision pipeline to analyze dashboard and screen rendering events for performance anomaly detection, leveraging PyTorch models deployed on AWS SageMaker inference endpoints.

- Built NLP-driven incident analysis services using LLMs (Azure OpenAI) and custom transformers to auto-summarize logs and traces, accelerating root cause analysis by 60% for on-call engineers.

- Designed and deployed scalable model serving infrastructure on Kubernetes with GPU-enabled nodes, utilizing Terraform and Helm for infrastructure-as-code, supporting over 50,000 inferences per second.

- Integrated machine learning models with existing monitoring systems via REST and gRPC APIs, enabling seamless correlation of AI-generated insights with traditional metrics and traces.

- Optimized AI model performance and inference latency by implementing model quantization, dynamic batching, and caching strategies within the Python-based serving layer.

- Collaborated with data science and platform teams to establish an MLOps pipeline using GitHub Actions, automating model training, validation, and canary deployment for computer vision and NLP services.

- Led the development of a distributed feature engineering platform in Python to preprocess logs and metrics, improving model training data quality and reducing feature computation time by 35%.

- Implemented comprehensive model monitoring and drift detection using Prometheus and Datadog custom metrics, ensuring high reliability and performance of production AI workflows.

- Mentored junior engineers on AI system design, cloud-based ML platforms, and distributed system patterns, fostering a culture of technical excellence and innovation. Microsoft Remote

Senior Software Engineer Jun 2021 – Jun 2023

- Architected enterprise AI platform services using Python, TensorFlow, and PyTorch to orchestrate LLM workflows and machine learning models across Azure and AWS SageMaker.

- Developed NLP-driven document intelligence services utilizing computer vision models for OCR and layout analysis, integrated into a vector search pipeline for enterprise knowledge retrieval.

- Built and optimized high-performance model inference APIs in Python (FastAPI) and C# (.NET Core), deployed on Azure Kubernetes Service (AKS) with auto-scaling and GPU acceleration.

- Designed a scalable feature store and model registry on Azure ML and AWS SageMaker to streamline the end-to- end ML lifecycle for cross-functional data science teams.

- Implemented machine learning pipelines for training and evaluating computer vision and NLP models, incorporating hyperparameter tuning and automated A/B testing frameworks.

- Engineered secure, low-latency APIs for real-time AI predictions, integrating with Azure Active Directory for identity and implementing comprehensive audit logging for compliance.

- Led the migration of legacy statistical modeling systems to modern deep learning architectures on TensorFlow/PyTorch, improving model accuracy by 25% on key classification tasks.

- Collaborated with applied scientists to productionize research models, focusing on model optimization techniques such as pruning, distillation, and ONNX conversion for edge deployment.

- Established ML monitoring and governance practices, tracking model performance, data drift, and business KPIs using Azure Monitor and custom dashboards.

- Provided technical leadership on AI system scalability, mentoring team members on distributed training, efficient data loading, and cloud-native deployment patterns.

Optum Basking Ridge, NJ

Senior Software Engineer Nov 2017 – May 2021

- Developed machine learning microservices in Python for healthcare claims analytics, leveraging scikit-learn and TensorFlow to predict claim adjudication outcomes and fraud patterns.

- Built and deployed NLP models for processing unstructured clinical notes and physician narratives, using spaCy and custom Keras models hosted on Azure Container Instances.

- Designed distributed data preprocessing pipelines for model training, handling high-volume, sensitive healthcare data securely within Azure cloud infrastructure.

- Integrated AI model predictions into core claims processing workflows via REST APIs, improving automated decision accuracy and reducing manual review volume by 30%.

- Implemented model versioning and A/B testing frameworks to evaluate new algorithms in production, using Azure ML for experiment tracking and management.

- Containerized ML models using Docker and orchestrated their deployment on Kubernetes, establishing blue-green deployment strategies for zero-downtime updates.

- Collaborated with data scientists to translate statistical models and prototypes into robust, scalable production services, focusing on latency, reliability, and compliance.

- Optimized batch inference pipelines for large-scale datasets, reducing job completion times by 40% through parallel processing and efficient resource allocation on Azure Databricks. Flatiron Health New York, NY

Software Engineer Jul 2014 - Oct 2017

- Built foundational data pipelines and ETL systems in Python and SQL to prepare oncology datasets for exploratory data analysis and early machine learning initiatives.

- Developed backend services to securely expose clinical data features for model training, implementing access controls and audit trails for compliant data usage.

- Assisted in the development of preliminary predictive models for patient outcomes using Python and scikit-learn, contributing to research on treatment efficacy.

- Engineered real-time data ingestion APIs that fed into analytics platforms, supporting the batch training cycles of early NLP models for clinical text processing.

- Gained foundational experience in statistical modeling and data quality practices essential for building reliable training datasets in a regulated healthcare environment.

EDUCATION

Princeton University NJ

Bachelor’s degree, Computer Science 2010 – 2014

Contact this candidate