Incubating ideas and creating technology, please stay tuned!

Head of AI leading a 35-person cross-functional global organization comprised of Machine Learning, Data, Software, and MLOps Engineers. Responsible for developing 0 to 1 AI B2B SaaS platform leveraging innovative multi-modal generative deep learning (LLMs) and automated knowledge graph construction. In 18 months:TL;DR• Hired and onboarded 5 principal team leads and 17 senior professionals• Built petabyte-scale event-driven data platform• Delivered 2 AI SaaS product launches supporting 10,000+ usersLeadership• Created detailed AI/Data strategic roadmap, aligned to product goals• Secured 8-figure incremental investment from CFO, responsible for budget and allocation• Led technical stakeholder engagement with customer executive leaders• Designed 3-phase organizational structure, to accommodate rapid scaling• Reviewed over 700 resumes, interviewed over 50 candidatesTechnical• Pioneered novel LLM strategies to create high-quality synthetic data • Orchestrated development of cloud-first MLOps platform on AWS• Architected and oversaw implementation of Kafka-based Kappa Architecture• Incorporated SoTA recommendation system design fusing knowledge graphs with vector stores (RAG)

MLOps & Platform Development:• Led the selection and implementation of a multi-cloud MLOps acceleration platform, evaluating market options, overseeing PoCs, and coordinating with a cross-functional team of SREs, Cloud Engineers, and ML practitioners for deployment on the company's private cloud.• Managed the internal roadmap for infrastructure optimization, with a focus on faster platform auto-scaling over spot/ephemeral compute, tighter integrations with Kafka and object storage, and multi-AZ GPU provisioning for efficient distributed training.• Designed and implemented standardized templates for production training and inference, streamlining project creation, improving deployment orchestration, and reducing MLE toil.• Established a "golden" ML Docker image system for secure, reproducible machine learning development and deployment.• Enabled an automated process for requesting large-scale training experiment compute requests, prioritizing trusted visibility/accountability.Architectural Decision-Making & Data Strategy:• Drafted numerous ADRs and key architectural guidance, impacting a wide range of decisions including the adoption of Kafka as our event-driven system, vendor selection for RDF graph database, and the choice of Kubeflow for ML model orchestration.• Drove data architecture strategies, including decisions related to Kappa architecture, data lake zone architecture, Lakehouse, data governance (covering lineage, provenance, and standardization), and the incorporation of data fabric principles with RDF.• Worked closely with product leadership to influence and guide the sequencing and roadmap of AI-enabled features.Leadership, Mentorship & Advocacy:• Mentored and coached mid/senior MLEs on conducting code and peer reviews on machine learning projects, fostering a culture of continuous learning and knowledge sharing.• Initiated and led an internal AI cohort for software engineers aspiring to transition into ML roles, cultivating talent growth.

Electronic Part Attribute Extraction:• Led the productionalization of an NLP inference graph for electronic part attribute extraction from manufacturer documents, architected in a variety of cloud platforms (Azure, AWS, and cloud-agnostic third parties).• Employed seven models including outlier-detection, classifiers, and deep learning RNNs to significantly enhance pipeline accuracy and performance.• Ensured efficiency and scalability with containerization, event-driven processing, auto-scaling, trivial parallelization, and high performance by integrating RAPIDS.• Solution saves over 50,000+ man hours annually.Advanced Model Training & Optimization:• Developed a Kubeflow continuous training pipeline for a subset of NLP models, focusing on code promotion for improved model governance.• Implemented Bayesian hyperparameter optimization that significantly reduced training time and elevated model quality.• Conceived a system for data version control for training/validation/test data, ensuring model reproducibility and result integrity.Forecasting System Redesign & Collaboration:• Redesigned and refactored an existing electronic parts price/supply chain forecasting system, using a mix of traditional and advanced methods (Holt-Winters, Prophet, RNN).• Collaborated with software engineers internationally, enhancing system efficiency by replacing expensive components with offline methods.Educational Initiatives & Leadership:• Developed and delivered data science and machine learning courses for division executives, product owners, and SMEs, educating over 500 individuals.• Mentored a group of software engineers in a company-wide machine learning hackathon that finished in the top 5%, imparting practical insights into data science project workflows.• Guided mid-level data engineers/scientists on efficient PySpark usage, and coached data analysts, product managers, and engineering managers on effective data visualization and dashboard creation.

• Generated $10MM+ incremental profit working on a team of data scientists to optimize pricing through experimentation and machine learning• Reduced purchasing project costs by 20% ($250K+) by creating interactive optimization model• Received one-on-one mentoring on communication and strategy from senior executives during a 6-month leadership rotation program • Crafted machine learning features, KPI’s, and business insights by exploring and modeling data related to fuel capacity, competitor landscape, demand seasonality, and customer geography• Designed and maintained internal data science team’s python library• Engaged with business leadership to hone a scientific mindset around measuring, experiments, and the use of statistics and machine learning