Building AI agents with multi-agent architectures.

Woven is a stealth startup backed by four top venture firms that is reimagining the way consumers travel. The company is building an AI-native Travel Concierge that simplifies the process of discovering, planning, and booking trips, all in one app. By pioneering breakthroughs in Agent Architectures and leveraging the latest advancements in Large Language Models (LLMs), Woven is poised to revolutionize the travel industry by helping consumers craft exquisite, personalized travel experiences.• Recruited and hired full-stack web, React Native, and AI / LLM team members.• Architected an LLM orchestration layer to asynchronously execute an ensemble of LLM based DAGs. Powered high quality personalization by leverage RAG workflows having long term memory, summarized contexts, and grounding techniques.• Created a novel git-backed system to produce many agent prompt configurations enabling LLM team to experiment, test, and systematically measure hundreds of prompt changes.• Helped built and ship a React Native application in the Apple app store.• Established LLM driven UI and clients where metadata was streamed so the client could incrementally enrich streaming text with richer UI components such as interactive hover cards and Google Maps.

Candidate Labs is a modern talent agency built as a technology company, pairing premier recruiters with technology and data to make the job search and hiring experience intelligent and seamless for candidates and employers.• Recruited and hired 8 full-stack engineers, and 7 PhD level data scientists / ML engineers.• Built pipeline that ingests 5+ terabytes of data across multiple data sources, applied ML inferences, and stored a people database of 100M+ candidates and 20M+ companies across horizontally scaled Postgres and Elastic Search clusters.• Created a recruiter friendly query language enabling candidate search matching employers top care-abouts to unique qualifications. The query itself was parsed and executed on a cluster of Panda data frames producing a bespoke and granular scoring model.

Built agile, collaborative and revenue generating UI/UX oriented teams at the nations largest issuer of personal loans. With an initial focus to boot-up a team charged to introduce a new technology (Node.js), over time grew to over forty five people spread across multiple teams with their own leadership across two countries. These teams transformed how engineering decisions were made by embracing a test and learn culture where we developed a discipline to only invest resources proportionate to quant based results. • Took a hands-on role to build a multi-touch attribution engine designed to power segmentation based personalization. Deployed a centralized service capable of supporting writes to a Cassandra cluster under some the highest volume loads, while also providing low latency read capabilities giving multiple sets of app servers real-time intelligence. Partnered directly with data science supplying them with a crisp data set allowing marketing to have a significantly higher visibility into spending.• Launched a complete redesign of company web presence, including re-imagining all top of funnel landing pages for loan originations. Through layering in A/B testing and experimentation team proved the refreshed UI / UX brand refreshed materially raised loan origination volume.• Tapped to create two tiger teams within personal loans, each to prove a new core business hypothesis that was thought to be a strong growth project. Started by quickly layering in an experiment which made small adjustments from UI to the credit model. Initial results increased origination volume so much an entire fully complimented team was layered in to double down.

Provided architecture evaluations to Fortune 500 companies helping them be successful in adopting Node.js. Expected to audit and map out entire system architecture above and below the Node.js stack, analyze gaps and opportunities for Node.js to add the highest value, then deliver a set of Node.js best practices to compliment a teams specific needs and organizational structure. Product managed NodeSource’s first product offering. Conducted product studies to identify market gaps within the Node.js ecosystem, ideated and defined scope. Entrusted to project manage a team of engineers by tactically setting priorities to execute the MVP. Worked with team members to champion the formation of a healthy engineering culture by bootstrapping a development process tuned to their strengths and emerging norms.

Managed two teams of expert level frontend (Node.js) and backend (Java) software engineers. Worked intently to promote a sense of teamwork and support which ultimately lead to having all major projects delivered on time. This resulted in a high sense of team unity and confidence. Fostering high team morale is the single greatest achievement in my career. Learned to understand team members’ strengths and how to complement an individuals’ weakness with skills from others. Tuned processes and workflow to fit the unique composition of the team. Promoted and implemented “divide and conquer” philosophy as a team. • Built based pub/sub messaging system over iFrames to allow multiple 3rd party embedded components (where BandPage "extensions" were the 3rd party) to communicate with each other acting as if they were within the same memory space as a single SPA• Developed a sophisticated affiliate tracking system, coordinating development milestones and collaborating with finance and product management to streamline user experience.• Created data feed powering integration enabling artists to sell merchandise on Spotify.

Was a key part of the team to help dramatically grow revenues during my tenure at Altrec. Setup an ESB / job management system to seamlessly integrate Altrec’s internal inventory systems used by the e-commerce website and warehouse ERP system together with Amazon’s Seller Central Marketplace which opened up double-digit revenue growth. Responsible for assigning tasks to appropriate resources. Created realistic timelines by collaborative pre-iteration planning in an agile environment. Directly managed offshore Java development team. Provided high-level architecture guidance by establishing clear priorities and expectations.

Contracted as an integral part of the development team assigned to enhance and repair much of the code base delivered from a previous outsourcing company. Took ownership of the implementation and deployment process in a highly client-centric environment. Used flat file to database conversions as well as data integrity validation.

Maintained and enhanced internal tools developed in a wide arrange of technologies to support an teams meet regulatory compliancy. Demonstrated ability to seamlessly support code built with little documentation and no hand off. Supported users by reverse engineering solutions through code analysis and researching functional requirements.Designed and deployed an ASP.Net AJAX enabled solution which automated the generation of Microsoft Word documents by creating a XLST on the fly. Process provided equivalent output to PDF stamping.

Part of a XP Agile development group responsible for creating and implementing a complete core banking system from the ground up. The project was built on an N-tiered proprietary application framework written in Java. A load balanced solution running in a multiple application/batch server environment. Developed JSP and JavaScript based presentation layer. Programmed in a heavy fat client architecture where state and optimistic offline locks had to be maintained within the JavaScript layer.