2023 UC Berkeley AI Summit Generative AI Hackathon Sees High Schoolers Outperform Professional Data Scientists for the Fourth Consecutive Year

Students and working professionals compete at the international event to create sophisticated enterprise-grade GenAI apps within 2 hours to solve real business problems

Aible, the fastest and safest path to enterprise generative AI, today announced the winners of the 2023 UC Berkeley AI Summit Generative AI Hackathon. The competition was part of the annual UC Berkeley AI Summit that aims to bridge the gap between AI research and business applications. The hackathon demonstrated that even high school students can create custom generative AI solutions with just 90 minutes of training, which is exactly in line with the US administration's goal of empowering a broader workforce with AI as articulated in the recent Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, where President Biden noted, “My Administration will seek to adapt job training and education to support a diverse workforce and help provide access to opportunities that AI creates.” 

This year’s GenAI hackathon had over 100 registrants with participation from students and working professionals from across 11 countries, including Austria, Canada, India, Indonesia, South Africa, UK, USA and others. Representative academic institutions included American High School, Arizona State University, Babson College, University of California Davis, University of California Berkeley, and Yeshiva University among others and enterprises included Broadcom, Ciena, Cisco, Electronic Arts, Johnson & Johnson, Meta, Optum, Wipro and more.

The hackathon event was judged by prominent names in the technology and AI space, including Gregory La Blanc, Distinguished Teaching Fellow at UC Berkeley, Haas School of Business; Tom H. Davenport, Distinguished Professor, Babson College; Dr. Matthew McCarville, CIO and Professor of Data Science and International Finance at University of Colorado and Former State Chief Data Officer at State of Florida; and Dr. Ali Arsanjani, Adjunct Professor at San Jose State University and the University of California, San Diego and Director, AI/ML Partner Engineering at Google.

This year’s hackathon challenged participants to create enterprise-grade and secure Generative AI apps using any tool of their choice, including Aible. For the fourth consecutive year, students with no prior data science experience beat expert AI professionals in terms of speed to create a fully functioning, scalable, and secure Generative AI solution. All of the winning submissions across different categories used Aible, continuing the trend from former competitions (2019 Challenge2021 Challenge and 2022 Challenge). The winners were all able to build enterprise generative apps in under two hours using Aible after receiving just 90 minutes of training. The fastest submission overall clocked in at just 16 minutes.

According to Dr. Matthew McCarville, a repeat judge at this contest, “Amazing to see a diverse group of participants from high school to data science and business professionals working on innovative things like prompt augmentation. Great to see the trend of high school students beating data scientists in both complexity and time is continuing, proof they are using their lack of experience as a way to lower their fear of the unknown when it comes to new technology!” 

Gregory La Blanc added, “Data science is evolving to the point where anyone can access capabilities that were once the domain of highly trained experts. The biggest obstacle for many is just the intimidation factor.”

Thomas H. Davenport said, "These skills are those that will make both students and professionals more productive and effective in their work. Everybody needs to learn them, and the hackathon winners are ahead of the game."

Dr. Arsanjani remarked, “It is heartwarming to see the uptake of generative AI by a spectrum of non-technical users as well as professionals, from high schoolers to business analysts and to see how they have used this hackathon as an opportunity to enrich their skills and demonstrate to the next generation who will be starting with generative AI and going beyond the initial novelty.”

Winning Generative AI Apps with Advanced Enterprise-Grade Features

The winning apps were fully functioning enterprise generative AI solutions that real-world business teams would be able to leverage to solve practical organizational challenges. The winning submissions were end-to-end automated generative AI solutions implemented in a secure Google Cloud instance, that leveraged VectorDB and Large Language Models integrations, specifically ChromaDB, Google’s PaLM 2 and OpenAI’s GPT-4.

Advanced Generative AI features were demonstrated by both students and professionals in the apps that they created after just a 90-minute live training session. 47% of the submitted entries showcased advanced techniques like prompt engineering and augmentation to make the GenAI responses higher quality and more meaningful, and 41% of submissions leveraged few-shot learning to further enhance the Large Language Model. Over 60% of the high school students were able to include advanced functionality, including few-shot learning and custom branding with only 90 minutes of training prior to the event. Similarly, 75% of the college students demonstrated their ability to build GenAI apps with prompt augmentation and few-shot learning functionality. The Aible platform massively reduces the level of skills necessary to leverage such advanced techniques - for example, students only had to edit the output of the GenAI response to make it better via few-shot learning.

$7,000 Awarded to Winners Across Four Categories

Seven participants were recognized as winners this year, including:

  • Noah Persily, a high schooler from Menlo School in California took home the overall top speed award using Aible to create a GenAI app, including an example of few-shot learning, in just 16 minutes. Bharath Gandoor from Cambrian Academy in San Jose, CA, took the award for the creativity category for high schoolers. Krish Patel from American High School in Fremont, CA, also received an award for completing the contest in half the allotted time.
  • Niranjan Kumar Kishore from Yeshiva University in New York took first place in the category of "College Student, Data Scientists or Comp Sci" while Tharun Prabhakar from Katz School of Science and Health in New York also received an award as he was just two minutes behind the winner.
  • Shabina Sherif, Sr. Product Manager from Optum Digital within UnitedHealth Group, won in the category of "Professionals - Business Users (not Data Scientists, Developers)"
  • Supriya Yoganandam, Data Engineer from Campbell Soup Company, took first place in the "Professionals - Data Scientists, Developers."

Aible Offers a Free Trial

Students and professionals looking to evaluate Aible for Enterprise Generative AI can now leverage the free trial for Aible.

About Aible

Aible is the fastest and safest path to Enterprise Generative AI. The business user solution brings end-to-end automation to GenAI and is implemented securely in private cloud accounts in 15 minutes. With sign-ups from over 1,200 enterprises, including 40+ Fortune 500 companies, the enterprise software that has a 5-star rating on Gartner Peer Insights enables business teams with advanced GenAI features to launch use cases including summarization, content generation, analytics and more on top of both structured and unstructured data. It is the only solution that significantly reduces wrong responses from GenAI with its ability to perform hallucination double-checks and deliver advanced features for few-shot learning, prompt augmentation, and more. The solution is built on an AI-first platform that helps teams find insights in minutes. Learn more at www.aible.com

Media Contact:

Email: pr@aible.com

Source: Aible