Brian Graham Email & Phone Number

New York, Ny, Us

• Leading development of large-scale generative AI features serving 4M+ members, including automated recipe import and food image recognition capabilities, while optimizing for peak winter diet season performance and scalability.• Driving significant CI/CD improvements through Python monorepo optimization, implementing service decoupling and incremental builds with uv that achieved 200% faster build times.

New York, Ny, Us

• Led WeightWatchers' generative AI strategy, managing both internal cross-functional engineering teams and Google-preferred vendor partnerships to prototype recipe chatbots and menu scanning features.• Made critical strategic pivot based on pilot study results, redirecting team focus from chatbot development to traditional value-adding tool feature in the app after identifying gaps in member AI readiness.• Successfully championed long-planned initiative to optimize growth funnels, directing implementation of contextual bandits for cancel-save offers that preserved $7M in customer lifetime value (LTV)

New York, Ny, Us

• Led development of Experience Engine, implementing ML models for predicting member disengagement risk and content affinity, while collaborating with product teams to integrate personalized content delivery infrastructure across the WeightWatchers platform• Successfully advocated for data science needs in company-wide Universal Analytics Service (UAS) modernization, ensuring new unified schema supported advanced ML model training requirements while collaborating across Android, iOS, and web platform teams• Modernized MLOps infrastructure by leading migration to Prefect from in-house workflow orchestration, establishing scalable experimentation platforms while maintaining production stability• Doubled data science team size to 12 members while fostering culture of ownership through structured onboarding, mentorship programs, and delegation of infrastructure responsibilities• Mentored two Columbia University masters student teams (8 students total) in developing generative AI prototypes for WW, teaching industry practices from version control to kubernetes deployment, resulting in successful demos of voice-based food tracking and AI recipe image generation at company-wide meeting

New York, Ny, Us

• Led development of an XGBoost-based ML model predicting key transitions in members' weight loss journeys, revolutionizing WeightWatchers' onboarding strategy and launching a multiyear initiative through cross-functional collaboration with product, design, and behavioral science teams• Championed early adoption of FastAPI to replace Flask across microservices, recognizing its integrated approach to API development with built-in Pydantic validation would eliminate dependency conflicts and reduce boilerplate compared to Flask's fragmented ecosystem of third-party packages• Architected separation of Business Logic and ML layers in service infrastructure, significantly improving A/B testing capabilities and reducing cross-team dependencies for model deployment• Developed and deployed an XGBoost classification model powering intelligent Invite-A-Friend content targeting, successfully pitching to stakeholders and shepherding from concept to production• Fostered cohesion across 12-person technical team during transition to remote-first environment, championing virtual collaboration practices and maintaining team culture through structured mentorship, pair programming, and regular social connections

New York, Ny, Us

• Architected and deployed a scalable Recipe Recommendation Engine using Flask, implementing production-grade API frameworks by integrating tooling for request validation, authentication, and response marshalling to enable team-wide microservice development• Developed a synthetic user behavior simulation system to enable QA testing of ML-powered features, generating realistic historical user data to validate personalized vs. non-personalized experiences in testing environments• Built an NLP pipeline using spaCy that parsed and structured complex food descriptions into dependency graphs, enabling the development of intuitive food-builder features that simplified tracking of customizable items like coffee and meals

San Francisco, Ca, Us

• Mentor PhD graduates and post-docs as they build prototypes of ML-powered data products over the course of a fast-paced 4 week program.• Provide both technical and product advice as they prepare to present their MVPs to hiring data teams across the industry.• Conduct technical mock interviews and guide fellows in navigating the interview process from preliminary screens through accepting an offer.

San Francisco, Ca, Us

• Built a mountain bike trail recommendation system “Rec & Ride” using a hybrid approach of content and collaborative recommender systems.• Scraped 3,000 trail pages for trail metadata and 1.5 million ridelogs for user-item interactions from Trailforks.com and engineered features using techniques including tf-idf text vectorization and SVD matrix factorization.• Deployed a web app on AWS EC2 built with Dash and Flask to deliver personalized ridingrecommendations to users based on their favorite trails and riding history.

Newark, De, Us

Investigated cartilage mechanics, specifically fluid transport within cartilage and how it relates to cartilage lubrication. My experiments combined the mechanical testing of cartilage explants with novel biomedical imaging techniques using laser scanning confocal microscopy. My research suggests that being active is critical to the transport of nutrients into cartilage and to keeping it hydrated to maintain its remarkably low friction.Collected and analyzed many types of data. I have developed algorithms in Matlab and Python to segment a variety of images, including 3D micro-CT images and multichannel 2D time series images from confocal microscopy, and to detect events in time series data from sensor signals. I have also used Matlab, Python, and R to perform statistical analyses and visualizations of this data for numerous presentations and publications.Designed and fabricated custom experimental test devices. I selected the actuators, sensors, power sources, and signal conditioners used in these devices, wired the electrical system, machined many of the metal components, and wrote significant portions of the control code in LabView. These devices had unique spatial constraints as they often had to interface with existing commercial imaging systems, which made this a difficult engineering challenge.

Kalamazoo, Mi, Us

Designed surgical tools for use with computer-aided knee replacement surgery and evaluated functional prototypes in a cadaver lab setting with a team of surgeons and engineers. Surveyed consulting surgeons and summarized their procedural preferences to form the guidelines for a standardized surgical workflow to aid in the design of new instrumentation.

Kalamazoo, Mi, Us

Combined surgeon and engineering inputs to design instruments for use in computer-aided knee replacement surgery. Rapidly iterated through prototypes using Pro/Engineer and additive manufacturing to quickly incorporate surgeon feedback and revise designs accordingly.

Responsible for the creation of tire shielding concepts through all stages of the development process -- from the whiteboard, to CAD model, to fabrication of the initial prototypes. Created the dual tire shield (DTS), which protects the sidewalls of tires on trailers pulled by military convoy vehicles from being punctured by bullets or debris in hostile and rugged environments. A patent for the DTS was granted in 2014.

Solon, Oh, Us

Designed a new geometry and assembly method for self-lubricating composite polymer/sheet metal bearings that are used in the automotive industry. The unique design, along with accompanying tooling, rivet, and pin, allows for the bearings to self-assemble during the riveting process, significantly reducing manufacturing time and costs by eliminating steps in the typical assembly process. The patent application for this invention received Notice of Allowance at the end of 2017.

Phd, Mechanical Engineering

Bs, Mechanical Engineering