Daylor Williams

University of California, Los Angeles

B.S. Cognitive Science, Computing Specialization · Los Angeles

Major GPA 3.90 · Cumulative GPA 3.72

AI Product Intern

UCLA Technology Development Group · Los Angeles

• Co-own product development for an enterprise agentic AI platform, translating invention-evaluation workflows into a system projected to reduce thousands of hours of manual prior-art research to approximately 15 minutes per disclosure.
• Run customer discovery with patent evaluators and licensing officers; map end-to-end workflows into user stories, prioritized capabilities, and product requirements that evolve as prototypes, testing, and stakeholder feedback expose new constraints.
• Define success metrics and lead iterative evaluation using golden datasets across sparse, crowded, and inconclusive technology fields; use benchmark failures and edge cases to refine retrieval strategy, agent behavior, and automation boundaries.
• Design and build multi-agent workflows in Python using REST APIs, OAuth, Azure, and European Patent Office infrastructure, preserving human judgment for high-stakes patentability decisions.
• Partner with business and engineering stakeholders on product architecture, interfaces, infrastructure, and deployment decisions while productionizing core services against a shared agent SDK.

Digital Assurance & Transparency Intern

PwC · San Francisco

• Led product design of an AI-enabled workpaper conversion solution: identified a high-cost workflow bottleneck, gathered user requirements, and delivered an agentic workflow with a modeled savings case of approximately 480 hours and $150K per engagement.
• Pitched the product and business case to the engagement partner, earning sponsorship for rollout planning as a scalable cost-reduction initiative across future assurance engagements.
• Independently built AI agents and internal dashboards beyond assigned engagement work to automate coordination, surface deadlines, and visualize deliverable status across team workflows.
• Supported assurance engagements for leading technology companies, evaluating digital products, services, data flows, and control environments affecting system reliability, data integrity, compliance, and assurance quality.

Machine Learning Research Assistant

Computational Vision & Learning Lab, UCLA · Los Angeles

• Lead first-author research targeting ICLR 2027 on representation engineering for DINOv3 vision transformers, investigating how foundation models encode object identity versus spatial relationships.
• Developed a representation decomposition framework using PCA, LDA, projection methods, and embedding-space analysis to isolate relational structure from object identity in high-dimensional representations.
• Designed and executed controlled experiments across 2,000+ synthetic images and 400 real-world objects spanning eight spatial relationships, demonstrating 97.9% cross-domain transfer.
• Design causal intervention experiments to manipulate internal representations and test whether relational knowledge can be predictably steered while ruling out low-level visual confounds.

GRAPES Search+ — Funding Discovery Product

UCLA Division of Graduate Education

• Independently conceived and built a 0-to-1 AI search product after identifying a usability gap through student conversations; scoped the MVP, designed the search experience, and built natural-language discovery over UCLA fellowship data.
• Audited 625 funding records and found only 350 remained active, then built a validation pipeline that surfaces live opportunities only; presented the product to Graduate Division leadership and am now advancing it toward implementation with the Assistant Dean of Fellowships and Financial Services.