San Francisco Bay Area
Machine Learning Leader with ~15 years of experience building and deploying AI systems in high-growth, mission-driven environments. Over 6+ years in leading and scaling ML teams from early-stage to 20+ across engineering and data science/machine learning. Proven track record delivering ML systems that integrate natural language, computer vision, and time series data. Hands-on technical contributor throughout leadership tenure, remaining 70%+ technical during early stage growth phases, stepping back to be more strategic as teams have grown. Deep expertise in MLOps, GCP/AWS/Azure, Kubernetes, model deployment, and cross-functional alignment with product, engineering, and research. Passionate about building transformative AI solutions to emerging tech.
• Led the scaling of ML organization from 4 to 20+ team members post-acquisition at Automation Anywhere, focusing on machine learning, data science, and engineering functions. • Directed the integration of Agentic solutions levering generative AI and advanced ML capabilities into an enterprise-grade automation platform, collaborating with product, UX, and engineering teams for end-to-end model deployment.
• Led integration of generative AI and advanced AI capabilities across our enterprise automation platform. • AI-focused projects driving the seamless adoption of machine learning solutions to enhance automation and user experience across the organization. • Manage the Process Discovery offering, leveraging state-of-the-art computer vision and AI technologies to help businesses uncover and optimize the workflows their users follow.
• Remained highly hands-on, contributing 70%+ of technical development in early leadership phases. Designed and implemented production-ready ML models using Python, TensorFlow, and PyTorch across NLP and CV domains. • Played a pivotal role in aligning technical strategy with fundraising milestones, prioritizing ML features to support customer retention and high-value sales deals during crucial growth periods. • Mentored and managed a high-performing team of ML engineers and researchers, fostering a collaborative, innovative culture rooted in continuous learning and experimentation.
• Increased model inference speed 1700% by creating a distributed system that leveraged Kubernetes and asynchronous messaging within google cloud • Improved clustering accuracy five-fold by fine-tuning a BERT transformer and CNN-based models to combine image and text information for prediction. • Worked constantly with data science and engineering leaders to drive the adoption of key project initiatives. • Steered data science team to prioritize sales and customer retention in high stakes times to drive additional funding rounds. • Automated airflow and elasticsearch deployments to development and production clusters
On-site Lead / Data Scientist / Machine Learning Engineer @ Apple Inc. • Lead a specialized data science team tasked with key machine learning projects, serving as a resource to assist members with individual projects and remain on track for client satisfaction • Applied machine learning algorithms to automatically find the best answer for customer issues drastically decreasing staff workload by 16,500 hours per month • Saved over 1000 hours per month by leveraging machine learning with NLP techniques to build a text classification algorithm that predicted with high accuracy whether advisor action was required • Applied deep learning (seq2seq) models to automate various use cases for translation and summarization • Slashed anomaly detection time for product issues by more than 10x by incorporating natural language processing, classification, and clustering within a Spark and a Hadoop ecosystem to scale to client needs
• Performed matched case-control study to identify the effectiveness of different treatment regimens for a specialty condition and provide better financial outcomes for patients • Claim Severity Predictions - Utilized machine learning techniques (gradient boosting and neural networks) to predict insurance claim severity by employing feature engineering, parameter optimization, and model ensembling