Engineering &
machine learning.

• Architect and lead engineer for recommendations at Saks Fifth Avenue, Neiman Marcus, and Bergdorf Goodman. The platform served 10M+ monthly users and 250M+ recommendations a month with 99.9% availability.
• Designed embedding-based candidate retrieval with real-time session features and learned ranking. Kept p95 serving latency below 300 ms; personalization increased conversion by 7%.
• Wrote Rust services for feature retrieval, caching, and model inference. Ran embedding and ranking models on autoscaled GPUs, and evaluated LLM-based retrieval and reranking through offline evaluation and A/B tests.
• Built batch and streaming pipelines processing more than 50 TB a month on AWS EKS. Added OpenTelemetry tracing, shared deployment components, and infrastructure as code, cutting production debugging time by about half.
• Managed the recommendation roadmap and infrastructure budget, including vendor selection and build-versus-buy decisions for models and GPU capacity.

PromotionsSenior Manager, Engineering → Director, Engineering

• Led data and engineering teams working on personalization, audience analysis, ad placement, and editorial automation.
• Developed reinforcement learning models for ad and content placement. Built an adaptive ad placement service that improved ad performance by 20%.
• Automated roughly 40% of manual editorial production with LLM and NLP tools while maintaining editorial quality.
• Built reusable ingestion pipelines and production APIs for audience segmentation, personalized content delivery, and advertising revenue analysis.

PromotionsDirector, Data Science → Senior Director, Machine Learning

• Led the team that built the first automated brand safety product, from prototype to production, generating $10M in new revenue.
• Built real-time video and text classification services with Kafka Streams and PyTorch, halving end-to-end classification latency.
• Rewrote performance-critical Python services in Rust, improving execution speed by 60% and cutting infrastructure costs by 40%.
• Ran 10+ services processing more than 1B decisions a month on AWS GPUs, using autoscaling and spot capacity to manage costs.

• Managed 15+ engineers and data scientists building ad inventory forecasts, optimization services, and data infrastructure.
• Developed reinforcement learning models for ad and content placement. Combined linear programming with demand forecasts to improve inventory management efficiency by 30%.
• Moved production workloads to Kubernetes and Airflow, replacing manual deployments with automated pipelines and canary releases.
• Built experimentation and deployment components adopted by other departments, and reported engineering plans, progress, and ROI to senior leadership.

PromotionsSenior Principal Data Scientist → Director, Data Science

• Implemented a computer vision and NLP pipeline with YOLO, ONNX, and spaCy to extract product, brand, and ingredient data from packaging, improving extraction accuracy by 30%.
• Automated extraction model updates with feedback loops, continuous retraining, and deployment pipelines.
• Containerized packaging extraction services on AWS, cutting compute costs by 25%.

PromotionsSenior Software Engineer, Machine Learning → Senior Data Scientist

• Led the team building retail recommendation, customer segmentation, forecasting, and marketing decision systems; marketing ROI improved by 10%.
• Built ETL pipelines and led a warehouse migration that cut manual data work by 40% and halved query times.
• Set analytics priorities with marketing, sales, and operations teams.

PromotionsAnalytics Specialist → Senior Analytics Specialist → Manager, Advanced Analytics

• Built data warehousing, IoT automation, and reporting systems, improving data retrieval speed by 35%.
• Automated IoT data collection and validation on AWS, halving manual data entry.

PromotionsBusiness Analyst → Senior Data Analyst → Program Manager