Profile picture

Allen Bolourchi

Part Time Lecturer of Technology and Applied Computing

Education

  • Other, Machine Learning Research Scientist, California Institute of Technology
  • Doctoral Degree, Evolutionary algorithms and Pareto Frontier optimization for predictive analysis of human spine and dynamical systems using regression trees with regularization, University of Southern California
  • Master's Degree, MBA in Tech/AI and Finance, University of California Los Angeles -Management School
  • Master's Degree, Electrical Engineering, University of Southern California

Biography

Allen Bolourchi is an Adjunct Professor of Generative AI and Natural Language Processing at the USC Viterbi School of Engineering and a senior leader of AI research and product management at Meta Superintelligence Labs (MSL), Fundamental AI Research (FAIR). His work spans frontier AI research, product, and research-to-product execution, with a focus on frontier models, agentic AI, world models, computer-use agents, multimodal intelligence, wearables, and smart glasses.


University of Southern California



At USC, Allen directs the Capstone of the Artificial Intelligence Minor, ITP 459: Generative AI and Natural Language Processing. He designed the course from the ground up as a highly hands-on program that combines the technical foundations of modern generative AI with AI product strategy and practical product development.


The course is structured around three core components:


(1) Generative AI & NLP Technical Foundations: Students develop a technical understanding of modern Generative AI and Natural Language Processing across 9 core areas, including transformers and attention, large language models, prompting and reasoning, fine-tuning, retrieval-augmented generation (RAG), vector databases, agentic AI systems, model evaluation, and AI safety.


(2) GenAI Case Project: Students analyze emerging AI companies from both product and market perspectives, often selecting startups from recent Y Combinator cohorts. Teams evaluate companies across 4 dimensions—AI product offerings, technical approaches, competitive positioning, and business models—and present their findings and strategic analysis to the class.


(3) AI Product Capstone: Students apply course concepts to build an AI product from scratch through a 5-stage 0-to-1 lifecycle: problem definition, product ideation, technical architecture, prototyping, and final demonstration. Students build working products using frontier-model APIs and AI-assisted coding tools, including Gemini, Claude Code, and Codex, gaining hands-on experience turning an AI concept into a functional product.


USC ITP 459: Generative AI and Natural Language Processing


Meta Superintelligence Labs and FAIR



At Meta Superintelligence Labs and FAIR, Allen leads AI research and product management spanning frontier models, agentic AI, world models, computer-use agents, multimodal systems, model evaluation, and research-to-product incubation. He develops multi-year AI research and research-to-product strategies across major FAIR research organizations, working with organizational leadership and senior research and engineering leaders to prioritize investments across research, compute, and product, and to translate emerging research into new AI product capabilities for end users. Allen leads the largest research organization within FAIR and incubated a new research organization from the ground up after joining MSL.


A central theme of his work is 0-to-1 AI research and product incubation: connecting frontier research with real-world products. He has helped establish new research initiatives, define technical and product direction, develop organizational strategies, and create evaluation and development systems that move new capabilities from early research toward deployment.


Allen has co-authored research with Yann LeCun and other FAIR researchers on world models, joint-embedding predictive architectures, multimodal intelligence, computer-use agents, video understanding, and proactive AI assistants. See the list of his publications below.


Apple



Previously, Allen was a Global Technical Product and Program Manager for AI/ML at Apple, focused on personalization, growth, and monetization for the App Store and Apple Arcade. He led the development of machine-learning capabilities spanning the customer lifecycle, from acquisition and engagement to retention, monetization, and win-back. His work included on-device and server-side AI models powering personalization and recommendations within the App Store and across marketing channels, including email, push notifications, and other customer touchpoints.


At Apple, Allen led cross-functional teams across machine learning, data science, engineering, analytics, and product to build and deploy omnichannel ML systems for a global user base. He also initiated and led the development of a five-year AI vision for Apple Marketing and Services, defining how generative AI, machine learning, analytics, and shared AI platforms could enhance customer journeys, content personalization, and existing AI capabilities across Apple's applications and marketing ecosystem.


Speaking



Allen has been a guest speaker at AI, technology, and business conferences and summits across the United States, including the MIT Sloan AI Conference at the MIT Media Lab, the GenAI Summit in Silicon Valley, the Future of Humanity Summit in Washington, D.C., and the Los Angeles Business Show at the Los Angeles Convention Center. His talks and presentations have focused on emerging AI technologies, generative AI, AI product development, and the evolution of frontier AI systems.


Education and Research



Allen holds a Ph.D. in Engineering with a focus on AI and Computational Intelligence from the University of Southern California, where he developed novel machine-learning and evolutionary optimization algorithms for predictive analysis of the human spine and dynamical systems, including Pareto-frontier optimization and regularized regression-tree methods.


He also holds an M.S. in Electrical Engineering from UCLA and an MBA from the UCLA Anderson School of Management, with a focus on AI and technology management and finance. Allen was also a Machine Learning Research Scientist at Caltech, conducting research in predictive analytics, stochastic systems, Bayesian methods, and statistical model selection using information criteria such as AIC and BIC.


Across academia and industry, Allen's career has centered on connecting frontier AI research, technical infrastructure, product strategy, and large-scale execution—from developing new machine-learning methods and shaping early research directions to building AI platforms and bringing emerging capabilities into frontier models, consumer products, and industry-first AI agents used at global scale.


Outside of Work



Outside of work, Allen is an avid runner and enjoys long-distance running. He has completed the Los Angeles Marathon three times and, as part of the Conquer LA Challenge, completed the Los Angeles Marathon, Rose Bowl/Pasadena Half Marathon, and Santa Monica 10K Classic. Running is one of his favorite ways to stay active and take on new personal challenges outside of work.


Recent Publications at Meta Superintelligence Labs (MSL), Fundamental AI Research (FAIR)


(1) VL-JEPA: Joint Embedding Predictive Architecture for Vision-language
(2) Planning with Reasoning using Vision Language World Model
(3) DigiData: Training and Evaluating General-Purpose Mobile Control Agents
(4) Computer Use at the Edge of the Statistical Precipice
(5) Action100M: A Large-scale Video Action Dataset
(6) Plan, Watch, Recover: A Benchmark and Architectures for Proactive Procedural Assistance

Appointments
  • Technology and Applied Computing Program
Office
  • Allen Bolourchi has not listed an office location.
Contact Information
  • allen.bolourchi@usc.edu
Links
Social Media